idnits 2.17.1 draft-ietf-p2psip-base-17.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- -- The document has examples using IPv4 documentation addresses according to RFC6890, but does not use any IPv6 documentation addresses. Maybe there should be IPv6 examples, too? Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year == Line 1863 has weird spacing: '...Options optio...' == Line 2117 has weird spacing: '...ionType type;...' == Line 2334 has weird spacing: '...tyValue ide...' == Line 2643 has weird spacing: '...ionType typ...' == Line 2645 has weird spacing: '...ionData val...' == (3 more instances...) -- The document seems to contain a disclaimer for pre-RFC5378 work, and may have content which was first submitted before 10 November 2008. The disclaimer is necessary when there are original authors that you have been unable to contact, or if some do not wish to grant the BCP78 rights to the IETF Trust. If you are able to get all authors (current and original) to grant those rights, you can and should remove the disclaimer; otherwise, the disclaimer is needed and you can ignore this comment. (See the Legal Provisions document at https://trustee.ietf.org/license-info for more information.) -- The document date (July 11, 2011) is 4672 days in the past. Is this intentional? -- Found something which looks like a code comment -- if you have code sections in the document, please surround them with '' and '' lines. Checking references for intended status: Proposed Standard ---------------------------------------------------------------------------- (See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) == Missing Reference: 'NodeIdLength' is mentioned on line 1758, but not defined -- Looks like a reference, but probably isn't: '0' on line 4645 == Missing Reference: 'RFC-AAAA' is mentioned on line 6357, but not defined ** Obsolete normative reference: RFC 2818 (Obsoleted by RFC 9110) ** Obsolete normative reference: RFC 3023 (Obsoleted by RFC 7303) ** Downref: Normative reference to an Informational RFC: RFC 3174 ** Obsolete normative reference: RFC 3447 (Obsoleted by RFC 8017) ** Obsolete normative reference: RFC 4347 (Obsoleted by RFC 6347) ** Obsolete normative reference: RFC 4395 (Obsoleted by RFC 7595) ** Obsolete normative reference: RFC 5245 (Obsoleted by RFC 8445, RFC 8839) ** Obsolete normative reference: RFC 5246 (Obsoleted by RFC 8446) ** Obsolete normative reference: RFC 5389 (Obsoleted by RFC 8489) ** Obsolete normative reference: RFC 5766 (Obsoleted by RFC 8656) ** Downref: Normative reference to an Informational RFC: RFC 6091 ** Downref: Normative reference to an Informational RFC: RFC 6234 == Outdated reference: A later version (-10) exists of draft-ietf-hip-reload-instance-03 == Outdated reference: A later version (-16) exists of draft-ietf-mmusic-ice-tcp-13 == Outdated reference: A later version (-15) exists of draft-ietf-p2psip-self-tuning-04 == Outdated reference: A later version (-15) exists of draft-ietf-p2psip-service-discovery-03 == Outdated reference: A later version (-21) exists of draft-ietf-p2psip-sip-06 -- Obsolete informational reference (is this intentional?): RFC 5201 (Obsoleted by RFC 7401) -- Obsolete informational reference (is this intentional?): RFC 5785 (Obsoleted by RFC 8615) Summary: 12 errors (**), 0 flaws (~~), 14 warnings (==), 7 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 P2PSIP C. Jennings 3 Internet-Draft Cisco 4 Intended status: Standards Track B. Lowekamp, Ed. 5 Expires: January 12, 2012 Skype 6 E. Rescorla 7 RTFM, Inc. 8 S. Baset 9 H. Schulzrinne 10 Columbia University 11 July 11, 2011 13 REsource LOcation And Discovery (RELOAD) Base Protocol 14 draft-ietf-p2psip-base-17 16 Abstract 18 This specification defines REsource LOcation And Discovery (RELOAD), 19 a peer-to-peer (P2P) signaling protocol for use on the Internet. A 20 P2P signaling protocol provides its clients with an abstract storage 21 and messaging service between a set of cooperating peers that form 22 the overlay network. RELOAD is designed to support a P2P Session 23 Initiation Protocol (P2PSIP) network, but can be utilized by other 24 applications with similar requirements by defining new usages that 25 specify the kinds of data that must be stored for a particular 26 application. RELOAD defines a security model based on a certificate 27 enrollment service that provides unique identities. NAT traversal is 28 a fundamental service of the protocol. RELOAD also allows access 29 from "client" nodes that do not need to route traffic or store data 30 for others. 32 Status of this Memo 34 This Internet-Draft is submitted in full conformance with the 35 provisions of BCP 78 and BCP 79. 37 Internet-Drafts are working documents of the Internet Engineering 38 Task Force (IETF). Note that other groups may also distribute 39 working documents as Internet-Drafts. The list of current Internet- 40 Drafts is at http://datatracker.ietf.org/drafts/current/. 42 Internet-Drafts are draft documents valid for a maximum of six months 43 and may be updated, replaced, or obsoleted by other documents at any 44 time. It is inappropriate to use Internet-Drafts as reference 45 material or to cite them other than as "work in progress." 47 This Internet-Draft will expire on January 12, 2012. 49 Copyright Notice 51 Copyright (c) 2011 IETF Trust and the persons identified as the 52 document authors. All rights reserved. 54 This document is subject to BCP 78 and the IETF Trust's Legal 55 Provisions Relating to IETF Documents 56 (http://trustee.ietf.org/license-info) in effect on the date of 57 publication of this document. Please review these documents 58 carefully, as they describe your rights and restrictions with respect 59 to this document. Code Components extracted from this document must 60 include Simplified BSD License text as described in Section 4.e of 61 the Trust Legal Provisions and are provided without warranty as 62 described in the Simplified BSD License. 64 This document may contain material from IETF Documents or IETF 65 Contributions published or made publicly available before November 66 10, 2008. The person(s) controlling the copyright in some of this 67 material may not have granted the IETF Trust the right to allow 68 modifications of such material outside the IETF Standards Process. 69 Without obtaining an adequate license from the person(s) controlling 70 the copyright in such materials, this document may not be modified 71 outside the IETF Standards Process, and derivative works of it may 72 not be created outside the IETF Standards Process, except to format 73 it for publication as an RFC or to translate it into languages other 74 than English. 76 Table of Contents 78 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 8 79 1.1. Basic Setting . . . . . . . . . . . . . . . . . . . . . 9 80 1.2. Architecture . . . . . . . . . . . . . . . . . . . . . . 10 81 1.2.1. Usage Layer . . . . . . . . . . . . . . . . . . . . 13 82 1.2.2. Message Transport . . . . . . . . . . . . . . . . . 14 83 1.2.3. Storage . . . . . . . . . . . . . . . . . . . . . . 14 84 1.2.4. Topology Plugin . . . . . . . . . . . . . . . . . . 15 85 1.2.5. Forwarding and Link Management Layer . . . . . . . . 15 86 1.3. Security . . . . . . . . . . . . . . . . . . . . . . . . 16 87 1.4. Structure of This Document . . . . . . . . . . . . . . . 17 88 2. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 17 89 3. Overlay Management Overview . . . . . . . . . . . . . . . . . 20 90 3.1. Security and Identification . . . . . . . . . . . . . . 20 91 3.1.1. Shared-Key Security . . . . . . . . . . . . . . . . 21 92 3.2. Clients . . . . . . . . . . . . . . . . . . . . . . . . 21 93 3.2.1. Client Routing . . . . . . . . . . . . . . . . . . . 22 94 3.2.2. Minimum Functionality Requirements for Clients . . . 23 95 3.3. Routing . . . . . . . . . . . . . . . . . . . . . . . . 23 96 3.4. Connectivity Management . . . . . . . . . . . . . . . . 26 97 3.5. Overlay Algorithm Support . . . . . . . . . . . . . . . 27 98 3.5.1. Support for Pluggable Overlay Algorithms . . . . . . 27 99 3.5.2. Joining, Leaving, and Maintenance Overview . . . . . 27 100 3.6. First-Time Setup . . . . . . . . . . . . . . . . . . . . 29 101 3.6.1. Initial Configuration . . . . . . . . . . . . . . . 29 102 3.6.2. Enrollment . . . . . . . . . . . . . . . . . . . . . 29 103 4. Application Support Overview . . . . . . . . . . . . . . . . 29 104 4.1. Data Storage . . . . . . . . . . . . . . . . . . . . . . 30 105 4.1.1. Storage Permissions . . . . . . . . . . . . . . . . 31 106 4.1.2. Replication . . . . . . . . . . . . . . . . . . . . 32 107 4.2. Usages . . . . . . . . . . . . . . . . . . . . . . . . . 32 108 4.3. Service Discovery . . . . . . . . . . . . . . . . . . . 33 109 4.4. Application Connectivity . . . . . . . . . . . . . . . . 33 110 5. Overlay Management Protocol . . . . . . . . . . . . . . . . . 33 111 5.1. Message Receipt and Forwarding . . . . . . . . . . . . . 34 112 5.1.1. Responsible ID . . . . . . . . . . . . . . . . . . . 34 113 5.1.2. Other ID . . . . . . . . . . . . . . . . . . . . . . 35 114 5.1.3. Private ID . . . . . . . . . . . . . . . . . . . . . 37 115 5.2. Symmetric Recursive Routing . . . . . . . . . . . . . . 37 116 5.2.1. Request Origination . . . . . . . . . . . . . . . . 37 117 5.2.2. Response Origination . . . . . . . . . . . . . . . . 38 118 5.3. Message Structure . . . . . . . . . . . . . . . . . . . 38 119 5.3.1. Presentation Language . . . . . . . . . . . . . . . 39 120 5.3.1.1. Common Definitions . . . . . . . . . . . . . . . 39 121 5.3.2. Forwarding Header . . . . . . . . . . . . . . . . . 42 122 5.3.2.1. Processing Configuration Sequence Numbers . . . . 44 123 5.3.2.2. Destination and Via Lists . . . . . . . . . . . . 45 124 5.3.2.3. Forwarding Options . . . . . . . . . . . . . . . 47 125 5.3.3. Message Contents Format . . . . . . . . . . . . . . 48 126 5.3.3.1. Response Codes and Response Errors . . . . . . . 49 127 5.3.4. Security Block . . . . . . . . . . . . . . . . . . . 51 128 5.4. Overlay Topology . . . . . . . . . . . . . . . . . . . . 55 129 5.4.1. Topology Plugin Requirements . . . . . . . . . . . . 55 130 5.4.2. Methods and types for use by topology plugins . . . 55 131 5.4.2.1. Join . . . . . . . . . . . . . . . . . . . . . . 55 132 5.4.2.2. Leave . . . . . . . . . . . . . . . . . . . . . . 56 133 5.4.2.3. Update . . . . . . . . . . . . . . . . . . . . . 57 134 5.4.2.4. RouteQuery . . . . . . . . . . . . . . . . . . . 57 135 5.4.2.5. Probe . . . . . . . . . . . . . . . . . . . . . . 58 136 5.5. Forwarding and Link Management Layer . . . . . . . . . . 60 137 5.5.1. Attach . . . . . . . . . . . . . . . . . . . . . . . 61 138 5.5.1.1. Request Definition . . . . . . . . . . . . . . . 61 139 5.5.1.2. Response Definition . . . . . . . . . . . . . . . 64 140 5.5.1.3. Using ICE With RELOAD . . . . . . . . . . . . . . 65 141 5.5.1.4. Collecting STUN Servers . . . . . . . . . . . . . 65 142 5.5.1.5. Gathering Candidates . . . . . . . . . . . . . . 66 143 5.5.1.6. Prioritizing Candidates . . . . . . . . . . . . . 66 144 5.5.1.7. Encoding the Attach Message . . . . . . . . . . . 67 145 5.5.1.8. Verifying ICE Support . . . . . . . . . . . . . . 67 146 5.5.1.9. Role Determination . . . . . . . . . . . . . . . 67 147 5.5.1.10. Full ICE . . . . . . . . . . . . . . . . . . . . 68 148 5.5.1.11. No-ICE . . . . . . . . . . . . . . . . . . . . . 68 149 5.5.1.12. Subsequent Offers and Answers . . . . . . . . . . 68 150 5.5.1.13. Sending Media . . . . . . . . . . . . . . . . . . 69 151 5.5.1.14. Receiving Media . . . . . . . . . . . . . . . . . 69 152 5.5.2. AppAttach . . . . . . . . . . . . . . . . . . . . . 69 153 5.5.2.1. Request Definition . . . . . . . . . . . . . . . 69 154 5.5.2.2. Response Definition . . . . . . . . . . . . . . . 70 155 5.5.3. Ping . . . . . . . . . . . . . . . . . . . . . . . . 70 156 5.5.3.1. Request Definition . . . . . . . . . . . . . . . 71 157 5.5.3.2. Response Definition . . . . . . . . . . . . . . . 71 158 5.5.4. ConfigUpdate . . . . . . . . . . . . . . . . . . . . 71 159 5.5.4.1. Request Definition . . . . . . . . . . . . . . . 72 160 5.5.4.2. Response Definition . . . . . . . . . . . . . . . 72 161 5.6. Overlay Link Layer . . . . . . . . . . . . . . . . . . . 73 162 5.6.1. Future Overlay Link Protocols . . . . . . . . . . . 74 163 5.6.1.1. HIP . . . . . . . . . . . . . . . . . . . . . . . 75 164 5.6.1.2. ICE-TCP . . . . . . . . . . . . . . . . . . . . . 75 165 5.6.1.3. Message-oriented Transports . . . . . . . . . . . 75 166 5.6.1.4. Tunneled Transports . . . . . . . . . . . . . . . 75 167 5.6.2. Framing Header . . . . . . . . . . . . . . . . . . . 76 168 5.6.3. Simple Reliability . . . . . . . . . . . . . . . . . 77 169 5.6.3.1. Retransmission and Flow Control . . . . . . . . . 78 170 5.6.4. DTLS/UDP with SR . . . . . . . . . . . . . . . . . . 79 171 5.6.5. TLS/TCP with FH, No-ICE . . . . . . . . . . . . . . 79 172 5.6.6. DTLS/UDP with SR, No-ICE . . . . . . . . . . . . . . 80 173 5.7. Fragmentation and Reassembly . . . . . . . . . . . . . . 80 174 6. Data Storage Protocol . . . . . . . . . . . . . . . . . . . . 81 175 6.1. Data Signature Computation . . . . . . . . . . . . . . . 82 176 6.2. Data Models . . . . . . . . . . . . . . . . . . . . . . 83 177 6.2.1. Single Value . . . . . . . . . . . . . . . . . . . . 84 178 6.2.2. Array . . . . . . . . . . . . . . . . . . . . . . . 84 179 6.2.3. Dictionary . . . . . . . . . . . . . . . . . . . . . 85 180 6.3. Access Control Policies . . . . . . . . . . . . . . . . 85 181 6.3.1. USER-MATCH . . . . . . . . . . . . . . . . . . . . . 86 182 6.3.2. NODE-MATCH . . . . . . . . . . . . . . . . . . . . . 86 183 6.3.3. USER-NODE-MATCH . . . . . . . . . . . . . . . . . . 86 184 6.3.4. NODE-MULTIPLE . . . . . . . . . . . . . . . . . . . 86 185 6.4. Data Storage Methods . . . . . . . . . . . . . . . . . . 87 186 6.4.1. Store . . . . . . . . . . . . . . . . . . . . . . . 87 187 6.4.1.1. Request Definition . . . . . . . . . . . . . . . 87 188 6.4.1.2. Response Definition . . . . . . . . . . . . . . . 91 189 6.4.1.3. Removing Values . . . . . . . . . . . . . . . . . 93 190 6.4.2. Fetch . . . . . . . . . . . . . . . . . . . . . . . 93 191 6.4.2.1. Request Definition . . . . . . . . . . . . . . . 94 192 6.4.2.2. Response Definition . . . . . . . . . . . . . . . 96 193 6.4.3. Stat . . . . . . . . . . . . . . . . . . . . . . . . 97 194 6.4.3.1. Request Definition . . . . . . . . . . . . . . . 97 195 6.4.3.2. Response Definition . . . . . . . . . . . . . . . 97 196 6.4.4. Find . . . . . . . . . . . . . . . . . . . . . . . . 99 197 6.4.4.1. Request Definition . . . . . . . . . . . . . . . 99 198 6.4.4.2. Response Definition . . . . . . . . . . . . . . . 100 199 6.4.5. Defining New Kinds . . . . . . . . . . . . . . . . . 101 200 7. Certificate Store Usage . . . . . . . . . . . . . . . . . . . 101 201 8. TURN Server Usage . . . . . . . . . . . . . . . . . . . . . . 102 202 9. Chord Algorithm . . . . . . . . . . . . . . . . . . . . . . . 104 203 9.1. Overview . . . . . . . . . . . . . . . . . . . . . . . . 105 204 9.2. Hash Function . . . . . . . . . . . . . . . . . . . . . 105 205 9.3. Routing . . . . . . . . . . . . . . . . . . . . . . . . 105 206 9.4. Redundancy . . . . . . . . . . . . . . . . . . . . . . . 106 207 9.5. Joining . . . . . . . . . . . . . . . . . . . . . . . . 106 208 9.6. Routing Attaches . . . . . . . . . . . . . . . . . . . . 107 209 9.7. Updates . . . . . . . . . . . . . . . . . . . . . . . . 107 210 9.7.1. Handling Neighbor Failures . . . . . . . . . . . . . 109 211 9.7.2. Handling Finger Table Entry Failure . . . . . . . . 110 212 9.7.3. Receiving Updates . . . . . . . . . . . . . . . . . 110 213 9.7.4. Stabilization . . . . . . . . . . . . . . . . . . . 111 214 9.7.4.1. Updating neighbor table . . . . . . . . . . . . . 111 215 9.7.4.2. Refreshing finger table . . . . . . . . . . . . . 111 216 9.7.4.3. Adjusting finger table size . . . . . . . . . . . 112 217 9.7.4.4. Detecting partitioning . . . . . . . . . . . . . 113 218 9.8. Route query . . . . . . . . . . . . . . . . . . . . . . 113 219 9.9. Leaving . . . . . . . . . . . . . . . . . . . . . . . . 114 221 10. Enrollment and Bootstrap . . . . . . . . . . . . . . . . . . 115 222 10.1. Overlay Configuration . . . . . . . . . . . . . . . . . 115 223 10.1.1. Relax NG Grammar . . . . . . . . . . . . . . . . . . 121 224 10.2. Discovery Through Configuration Server . . . . . . . . . 123 225 10.3. Credentials . . . . . . . . . . . . . . . . . . . . . . 124 226 10.3.1. Self-Generated Credentials . . . . . . . . . . . . . 125 227 10.4. Searching for a Bootstrap Node . . . . . . . . . . . . . 126 228 10.5. Contacting a Bootstrap Node . . . . . . . . . . . . . . 126 229 11. Message Flow Example . . . . . . . . . . . . . . . . . . . . 127 230 12. Security Considerations . . . . . . . . . . . . . . . . . . . 133 231 12.1. Overview . . . . . . . . . . . . . . . . . . . . . . . . 133 232 12.2. Attacks on P2P Overlays . . . . . . . . . . . . . . . . 134 233 12.3. Certificate-based Security . . . . . . . . . . . . . . . 134 234 12.4. Shared-Secret Security . . . . . . . . . . . . . . . . . 135 235 12.5. Storage Security . . . . . . . . . . . . . . . . . . . . 136 236 12.5.1. Authorization . . . . . . . . . . . . . . . . . . . 136 237 12.5.2. Distributed Quota . . . . . . . . . . . . . . . . . 137 238 12.5.3. Correctness . . . . . . . . . . . . . . . . . . . . 137 239 12.5.4. Residual Attacks . . . . . . . . . . . . . . . . . . 137 240 12.6. Routing Security . . . . . . . . . . . . . . . . . . . . 138 241 12.6.1. Background . . . . . . . . . . . . . . . . . . . . . 138 242 12.6.2. Admissions Control . . . . . . . . . . . . . . . . . 139 243 12.6.3. Peer Identification and Authentication . . . . . . . 139 244 12.6.4. Protecting the Signaling . . . . . . . . . . . . . . 140 245 12.6.5. Residual Attacks . . . . . . . . . . . . . . . . . . 140 246 13. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 141 247 13.1. Well-Known URI Registration . . . . . . . . . . . . . . 141 248 13.2. Port Registrations . . . . . . . . . . . . . . . . . . . 141 249 13.3. Overlay Algorithm Types . . . . . . . . . . . . . . . . 142 250 13.4. Access Control Policies . . . . . . . . . . . . . . . . 142 251 13.5. Application-ID . . . . . . . . . . . . . . . . . . . . . 142 252 13.6. Data Kind-ID . . . . . . . . . . . . . . . . . . . . . . 143 253 13.7. Data Model . . . . . . . . . . . . . . . . . . . . . . . 143 254 13.8. Message Codes . . . . . . . . . . . . . . . . . . . . . 143 255 13.9. Error Codes . . . . . . . . . . . . . . . . . . . . . . 144 256 13.10. Overlay Link Types . . . . . . . . . . . . . . . . . . . 145 257 13.11. Overlay Link Protocols . . . . . . . . . . . . . . . . . 145 258 13.12. Forwarding Options . . . . . . . . . . . . . . . . . . . 146 259 13.13. Probe Information Types . . . . . . . . . . . . . . . . 146 260 13.14. Message Extensions . . . . . . . . . . . . . . . . . . . 146 261 13.15. reload URI Scheme . . . . . . . . . . . . . . . . . . . 147 262 13.15.1. URI Registration . . . . . . . . . . . . . . . . . . 147 263 13.16. Media Type Registration . . . . . . . . . . . . . . . . 148 264 14. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 149 265 15. References . . . . . . . . . . . . . . . . . . . . . . . . . 150 266 15.1. Normative References . . . . . . . . . . . . . . . . . . 150 267 15.2. Informative References . . . . . . . . . . . . . . . . . 151 268 Appendix A. Routing Alternatives . . . . . . . . . . . . . . . . 154 269 A.1. Iterative vs Recursive . . . . . . . . . . . . . . . . . 155 270 A.2. Symmetric vs Forward response . . . . . . . . . . . . . 155 271 A.3. Direct Response . . . . . . . . . . . . . . . . . . . . 155 272 A.4. Relay Peers . . . . . . . . . . . . . . . . . . . . . . 157 273 A.5. Symmetric Route Stability . . . . . . . . . . . . . . . 157 274 Appendix B. Why Clients? . . . . . . . . . . . . . . . . . . . . 158 275 B.1. Why Not Only Peers? . . . . . . . . . . . . . . . . . . 158 276 B.2. Clients as Application-Level Agents . . . . . . . . . . 158 277 Appendix C. Change Log . . . . . . . . . . . . . . . . . . . . . 159 278 C.1. Changes since draft-ietf-p2psip-reload-13 . . . . . . . 159 279 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 159 281 1. Introduction 283 This document defines REsource LOcation And Discovery (RELOAD), a 284 peer-to-peer (P2P) signaling protocol for use on the Internet. It 285 provides a generic, self-organizing overlay network service, allowing 286 nodes to efficiently route messages to other nodes and to efficiently 287 store and retrieve data in the overlay. RELOAD provides several 288 features that are critical for a successful P2P protocol for the 289 Internet: 291 Security Framework: A P2P network will often be established among a 292 set of peers that do not trust each other. RELOAD leverages a 293 central enrollment server to provide credentials for each peer 294 which can then be used to authenticate each operation. This 295 greatly reduces the possible attack surface. 297 Usage Model: RELOAD is designed to support a variety of 298 applications, including P2P multimedia communications with the 299 Session Initiation Protocol [I-D.ietf-p2psip-sip]. RELOAD allows 300 the definition of new application usages, each of which can define 301 its own data types, along with the rules for their use. This 302 allows RELOAD to be used with new applications through a simple 303 documentation process that supplies the details for each 304 application. 306 NAT Traversal: RELOAD is designed to function in environments where 307 many if not most of the nodes are behind NATs or firewalls. 308 Operations for NAT traversal are part of the base design, 309 including using ICE to establish new RELOAD or application 310 protocol connections. 312 High Performance Routing: The very nature of overlay algorithms 313 introduces a requirement that peers participating in the P2P 314 network route requests on behalf of other peers in the network. 315 This introduces a load on those other peers, in the form of 316 bandwidth and processing power. RELOAD has been defined with a 317 simple, lightweight forwarding header, thus minimizing the amount 318 of effort required by intermediate peers. 320 Pluggable Overlay Algorithms: RELOAD has been designed with an 321 abstract interface to the overlay layer to simplify implementing a 322 variety of structured (e.g., distributed hash tables) and 323 unstructured overlay algorithms. This specification also defines 324 how RELOAD is used with the Chord based DHT algorithm, which is 325 mandatory to implement. Specifying a default "must implement" 326 overlay algorithm promotes interoperability, while extensibility 327 allows selection of overlay algorithms optimized for a particular 328 application. 330 These properties were designed specifically to meet the requirements 331 for a P2P protocol to support SIP. This document defines the base 332 protocol for the distributed storage and location service, as well as 333 critical usages for NAT traversal and security. The SIP Usage itself 334 is described separately in [I-D.ietf-p2psip-sip]. RELOAD is not 335 limited to usage by SIP and could serve as a tool for supporting 336 other P2P applications with similar needs. 338 1.1. Basic Setting 340 In this section, we provide a brief overview of the operational 341 setting for RELOAD. A RELOAD Overlay Instance consists of a set of 342 nodes arranged in a connected graph. Each node in the overlay is 343 assigned a numeric Node-ID which, together with the specific overlay 344 algorithm in use, determines its position in the graph and the set of 345 nodes it connects to. The figure below shows a trivial example which 346 isn't drawn from any particular overlay algorithm, but was chosen for 347 convenience of representation. 349 +--------+ +--------+ +--------+ 350 | Node 10|--------------| Node 20|--------------| Node 30| 351 +--------+ +--------+ +--------+ 352 | | | 353 | | | 354 +--------+ +--------+ +--------+ 355 | Node 40|--------------| Node 50|--------------| Node 60| 356 +--------+ +--------+ +--------+ 357 | | | 358 | | | 359 +--------+ +--------+ +--------+ 360 | Node 70|--------------| Node 80|--------------| Node 90| 361 +--------+ +--------+ +--------+ 362 | 363 | 364 +--------+ 365 | Node 85| 366 |(Client)| 367 +--------+ 369 Because the graph is not fully connected, when a node wants to send a 370 message to another node, it may need to route it through the network. 371 For instance, Node 10 can talk directly to nodes 20 and 40, but not 372 to Node 70. In order to send a message to Node 70, it would first 373 send it to Node 40 with instructions to pass it along to Node 70. 374 Different overlay algorithms will have different connectivity graphs, 375 but the general idea behind all of them is to allow any node in the 376 graph to efficiently reach every other node within a small number of 377 hops. 379 The RELOAD network is not only a messaging network. It is also a 380 storage network. Records are stored under numeric addresses which 381 occupy the same space as node identifiers. Peers are responsible for 382 storing the data associated with some set of addresses as determined 383 by their Node-ID. For instance, we might say that every peer is 384 responsible for storing any data value which has an address less than 385 or equal to its own Node-ID, but greater than the next lowest 386 Node-ID. Thus, Node-20 would be responsible for storing values 387 11-20. 389 RELOAD also supports clients. These are nodes which have Node-IDs 390 but do not participate in routing or storage. For instance, in the 391 figure above Node 85 is a client. It can route to the rest of the 392 RELOAD network via Node 80, but no other node will route through it 393 and Node 90 is still responsible for all addresses between 81-90. We 394 refer to non-client nodes as peers. 396 Other applications (for instance, SIP) can be defined on top of 397 RELOAD and use these two basic RELOAD services to provide their own 398 services. 400 1.2. Architecture 402 RELOAD is fundamentally an overlay network. The following figure 403 shows the layered RELOAD architecture. 405 Application 407 +-------+ +-------+ 408 | SIP | | XMPP | ... 409 | Usage | | Usage | 410 +-------+ +-------+ 411 ------------------------------------ Messaging Service Boundary 412 +------------------+ +---------+ 413 | Message |<--->| Storage | 414 | Transport | +---------+ 415 +------------------+ ^ 416 ^ ^ | 417 | v v 418 | +-------------------+ 419 | | Topology | 420 | | Plugin | 421 | +-------------------+ 422 | ^ 423 v v 424 +------------------+ 425 | Forwarding & | 426 | Link Management | 427 +------------------+ 428 ------------------------------------ Overlay Link Service Boundary 429 +-------+ +------+ 430 |TLS | |DTLS | ... 431 +-------+ +------+ 433 The major components of RELOAD are: 435 Usage Layer: Each application defines a RELOAD usage; a set of data 436 kinds and behaviors which describe how to use the services 437 provided by RELOAD. These usages all talk to RELOAD through a 438 common Message Transport Service. 440 Message Transport: Handles end-to-end reliability, manages request 441 state for the usages, and forwards Store and Fetch operations to 442 the Storage component. Delivers message responses to the 443 component initiating the request. 445 Storage: The Storage component is responsible for processing 446 messages relating to the storage and retrieval of data. It talks 447 directly to the Topology Plugin to manage data replication and 448 migration, and it talks to the Message Transport component to send 449 and receive messages. 451 Topology Plugin: The Topology Plugin is responsible for implementing 452 the specific overlay algorithm being used. It uses the Message 453 Transport component to send and receive overlay management 454 messages, to the Storage component to manage data replication, and 455 directly to the Forwarding Layer to control hop-by-hop message 456 forwarding. This component closely parallels conventional routing 457 algorithms, but is more tightly coupled to the Forwarding Layer 458 because there is no single "routing table" equivalent used by all 459 overlay algorithms. 461 Forwarding and Link Management Layer: Stores and implements the 462 routing table by providing packet forwarding services between 463 nodes. It also handles establishing new links between nodes, 464 including setting up connections across NATs using ICE. 466 Overlay Link Layer: Responsible for actually transporting traffic 467 directly between nodes. Each such protocol includes the 468 appropriate provisions for per-hop framing or hop-by-hop ACKs 469 required by unreliable transports. TLS [RFC5246] and DTLS 470 [RFC4347] are the currently defined "link layer" protocols used by 471 RELOAD for hop-by-hop communication. New protocols can be 472 defined, as described in Section 5.6.1 and Section 10.1. As this 473 document defines only TLS and DTLS, we use those terms throughout 474 the remainder of the document with the understanding that some 475 future specification may add new overlay link layers. 477 To further clarify the roles of the various layers, this figure 478 parallels the architecture with each layer's role from an overlay 479 perspective and implementation layer in the internet: 481 | Internet Model | 482 Real | Equivalent | Reload 483 Internet | in Overlay | Architecture 484 -------------+-----------------+------------------------------------ 485 | | +-------+ +-------+ 486 | Application | | SIP | | XMPP | ... 487 | | | Usage | | Usage | 488 | | +-------+ +-------+ 489 | | ---------------------------------- 490 | |+------------------+ +---------+ 491 | Transport || Message |<--->| Storage | 492 | || Transport | +---------+ 493 | |+------------------+ ^ 494 | | ^ ^ | 495 | | | v v 496 Application | | | +-------------------+ 497 | (Routing) | | | Topology | 498 | | | | Plugin | 499 | | | +-------------------+ 500 | | | ^ 501 | | v v 502 | Network | +------------------+ 503 | | | Forwarding & | 504 | | | Link Management | 505 | | +------------------+ 506 | | ---------------------------------- 507 Transport | Link | +-------+ +------+ 508 | | |TLS | |DTLS | ... 509 | | +-------+ +------+ 510 -------------+-----------------+------------------------------------ 511 Network | 512 | 513 Link | 515 1.2.1. Usage Layer 517 The top layer, called the Usage Layer, has application usages, such 518 as the SIP Registration Usage [I-D.ietf-p2psip-sip], that use the 519 abstract Message Transport Service provided by RELOAD. The goal of 520 this layer is to implement application-specific usages of the generic 521 overlay services provided by RELOAD. The usage defines how a 522 specific application maps its data into something that can be stored 523 in the overlay, where to store the data, how to secure the data, and 524 finally how applications can retrieve and use the data. 526 The architecture diagram shows both a SIP usage and an XMPP usage. A 527 single application may require multiple usages; for example a 528 softphone application may also require a voicemail usage. A usage 529 may define multiple kinds of data that are stored in the overlay and 530 may also rely on kinds originally defined by other usages. 532 Because the security and storage policies for each kind are dictated 533 by the usage defining the kind, the usages may be coupled with the 534 Storage component to provide security policy enforcement and to 535 implement appropriate storage strategies according to the needs of 536 the usage. The exact implementation of such an interface is outside 537 the scope of this specification. 539 1.2.2. Message Transport 541 The Message Transport component provides a generic message routing 542 service for the overlay. The Message Transport layer is responsible 543 for end-to-end message transactions, including retransmissions. Each 544 peer is identified by its location in the overlay as determined by 545 its Node-ID. A component that is a client of the Message Transport 546 can perform two basic functions: 548 o Send a message to a given peer specified by Node-ID or to the peer 549 responsible for a particular Resource-ID. 550 o Receive messages that other peers sent to a Node-ID or Resource-ID 551 for which the receiving peer is responsible. 553 All usages rely on the Message Transport component to send and 554 receive messages from peers. For instance, when a usage wants to 555 store data, it does so by sending Store requests. Note that the 556 Storage component and the Topology Plugin are themselves clients of 557 the Message Transport, because they need to send and receive messages 558 from other peers. 560 The Message Transport Service is similar to those described as 561 providing "Key based routing" (KBR), although as RELOAD supports 562 different overlay algorithms (including non-DHT overlay algorithms) 563 that calculate keys in different ways, the actual interface must 564 accept Resource Names rather than actual keys. 566 1.2.3. Storage 568 One of the major functions of RELOAD is to allow nodes to store data 569 in the overlay and to retrieve data stored by other nodes or by 570 themselves. The Storage component is responsible for processing data 571 storage and retrieval messages. For instance, the Storage component 572 might receive a Store request for a given resource from the Message 573 Transport. It would then query the appropriate usage before storing 574 the data value(s) in its local data store and sending a response to 575 the Message Transport for delivery to the requesting node. 576 Typically, these messages will come from other nodes, but depending 577 on the overlay topology, a node might be responsible for storing data 578 for itself as well, especially if the overlay is small. 580 A peer's Node-ID determines the set of resources that it will be 581 responsible for storing. However, the exact mapping between these is 582 determined by the overlay algorithm in use. The Storage component 583 will only receive a Store request from the Message Transport if this 584 peer is responsible for that Resource-ID. The Storage component is 585 notified by the Topology Plugin when the Resource-IDs for which it is 586 responsible change, and the Storage component is then responsible for 587 migrating resources to other peers, as required. 589 1.2.4. Topology Plugin 591 RELOAD is explicitly designed to work with a variety of overlay 592 algorithms. In order to facilitate this, the overlay algorithm 593 implementation is provided by a Topology Plugin so that each overlay 594 can select an appropriate overlay algorithm that relies on the common 595 RELOAD core protocols and code. 597 The Topology Plugin is responsible for maintaining the overlay 598 algorithm Routing Table, which is consulted by the Forwarding and 599 Link Management Layer before routing a message. When connections are 600 made or broken, the Forwarding and Link Management Layer notifies the 601 Topology Plugin, which adjusts the routing table as appropriate. The 602 Topology Plugin will also instruct the Forwarding and Link Management 603 Layer to form new connections as dictated by the requirements of the 604 overlay algorithm Topology. The Topology Plugin issues periodic 605 update requests through Message Transport to maintain and update its 606 Routing Table. 608 As peers enter and leave, resources may be stored on different peers, 609 so the Topology Plugin also keeps track of which peers are 610 responsible for which resources. As peers join and leave, the 611 Topology Plugin instructs the Storage component to issue resource 612 migration requests as appropriate, in order to ensure that other 613 peers have whatever resources they are now responsible for. The 614 Topology Plugin is also responsible for providing for redundant data 615 storage to protect against loss of information in the event of a peer 616 failure and to protect against compromised or subversive peers. 618 1.2.5. Forwarding and Link Management Layer 620 The Forwarding and Link Management Layer is responsible for getting a 621 message to the next peer, as determined by the Topology Plugin. This 622 Layer establishes and maintains the network connections as required 623 by the Topology Plugin. This layer is also responsible for setting 624 up connections to other peers through NATs and firewalls using ICE, 625 and it can elect to forward traffic using relays for NAT and firewall 626 traversal. 628 This layer provides a generic interface that allows the topology 629 plugin to control the overlay and resource operations and messages. 630 Since each overlay algorithm is defined and functions differently, we 631 generically refer to the table of other peers that the overlay 632 algorithm maintains and uses to route requests (neighbors) as a 633 Routing Table. The Topology Plugin actually owns the Routing Table, 634 and forwarding decisions are made by querying the Topology Plugin for 635 the next hop for a particular Node-ID or Resource-ID. If this node 636 is the destination of the message, the message is delivered to the 637 Message Transport. 639 This layer also utilizes a framing header to encapsulate messages as 640 they are forwarding along each hop. This header aids reliability 641 congestion control, flow control, etc. It has meaning only in the 642 context of that individual link. 644 The Forwarding and Link Management Layer sits on top of the Overlay 645 Link Layer protocols that carry the actual traffic. This 646 specification defines how to use DTLS and TLS protocols to carry 647 RELOAD messages. 649 1.3. Security 651 RELOAD's security model is based on each node having one or more 652 public key certificates. In general, these certificates will be 653 assigned by a central server which also assigns Node-IDs, although 654 self-signed certificates can be used in closed networks. These 655 credentials can be leveraged to provide communications security for 656 RELOAD messages. RELOAD provides communications security at three 657 levels: 659 Connection Level: Connections between peers are secured with TLS, 660 DTLS, or potentially some to be defined future protocol. 661 Message Level: Each RELOAD message must be signed. 662 Object Level: Stored objects must be signed by the creating peer. 664 These three levels of security work together to allow peers to verify 665 the origin and correctness of data they receive from other peers, 666 even in the face of malicious activity by other peers in the overlay. 667 RELOAD also provides access control built on top of these 668 communications security features. Because the peer responsible for 669 storing a piece of data can validate the signature on the data being 670 stored, the responsible peer can determine whether a given operation 671 is permitted or not. 673 RELOAD also provides an optional shared secret based admission 674 control feature using shared secrets and TLS-PSK. In order to form a 675 TLS connection to any node in the overlay, a new node needs to know 676 the shared overlay key, thus restricting access to authorized users 677 only. This feature is used together with certificate-based access 678 control, not as a replacement for it. It is typically used when 679 self-signed certificates are being used but would generally not be 680 used when the certificates were all signed by an enrollment server. 682 1.4. Structure of This Document 684 The remainder of this document is structured as follows. 686 o Section 2 provides definitions of terms used in this document. 687 o Section 3 provides an overview of the mechanisms used to establish 688 and maintain the overlay. 689 o Section 4 provides an overview of the mechanism RELOAD provides to 690 support other applications. 691 o Section 5 defines the protocol messages that RELOAD uses to 692 establish and maintain the overlay. 693 o Section 6 defines the protocol messages that are used to store and 694 retrieve data using RELOAD. 695 o Section 7 defines the Certificate Store Usage that is fundamental 696 to RELOAD security. 697 o Section 8 defines the TURN Server Usage needed to locate TURN 698 servers for NAT traversal. 699 o Section 9 defines a specific Topology Plugin using Chord based 700 algorithm. 701 o Section 10 defines the mechanisms that new RELOAD nodes use to 702 join the overlay for the first time. 703 o Section 11 provides an extended example. 705 2. Terminology 707 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 708 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 709 document are to be interpreted as described in RFC 2119 [RFC2119]. 711 Terms used in this document are defined inline when used and are also 712 defined below for reference. 714 DHT: A distributed hash table. A DHT is an abstract hash table 715 service realized by storing the contents of the hash table across 716 a set of peers. 718 Overlay Algorithm: An overlay algorithm defines the rules for 719 determining which peers in an overlay store a particular piece of 720 data and for determining a topology of interconnections amongst 721 peers in order to find a piece of data. 723 Overlay Instance: A specific overlay algorithm and the collection of 724 peers that are collaborating to provide read and write access to 725 it. There can be any number of overlay instances running in an IP 726 network at a time, and each operates in isolation of the others. 728 Peer: A host that is participating in the overlay. Peers are 729 responsible for holding some portion of the data that has been 730 stored in the overlay and also route messages on behalf of other 731 hosts as required by the Overlay Algorithm. 733 Client: A host that is able to store data in and retrieve data from 734 the overlay but which is not participating in routing or data 735 storage for the overlay. 737 Kind: A kind defines a particular type of data that can be stored in 738 the overlay. Applications define new Kinds to store the data they 739 use. Each Kind is identified with a unique integer called a 740 Kind-ID. 742 Node: We use the term "Node" to refer to a host that may be either a 743 Peer or a Client. Because RELOAD uses the same protocol for both 744 clients and peers, much of the text applies equally to both. 745 Therefore we use "Node" when the text applies to both Clients and 746 Peers and the more specific term (i.e. client or peer) when the 747 text applies only to Clients or only to Peers. 749 Node-ID: A fixed-length value that uniquely identifies a node. 750 Node-IDs of all 0s and all 1s are reserved and are invalid Node- 751 IDs. A value of zero is not used in the wire protocol but can be 752 used to indicate an invalid node in implementations and APIs. The 753 Node-ID of all 1s is used on the wire protocol as a wildcard. 755 Joining Peer: A node that is attempting to become a Peer in a 756 particular Overlay.. 758 Admitting Peer: A Peer in the Overlay which helps the Joining Peer 759 join the Overlay. 761 Bootstrap Node: A network node used by Joining Peers to help Nlocate 762 the Admitting Peer. 764 Peer Admission: The act of admitting a peer (the "Joining Peer" ) 765 into an Overlay. After the admission process is over, the joining 766 peer is a fully-functional peer of the overlay. During the 767 admission process, the joining peer may need to present 768 credentials to prove that it has sufficient authority to join the 769 overlay. 771 Resource: An object or group of objects associated with a string 772 identifier. See "Resource Name" below. 774 Resource Name: The potentially human readable name by which a 775 resource is identified. In unstructured P2P networks, the 776 resource name is sometimes used directly as a Resource-ID. In 777 structured P2P networks the resource name is typically mapped into 778 a Resource-ID by using the string as the input to hash function. 779 Structured and unstructured P2P networks are described in 780 [RFC5694]. A SIP resource, for example, is often identified by 781 its AOR which is an example of a Resource Name. 783 Resource-ID: A value that identifies some resources and which is 784 used as a key for storing and retrieving the resource. Often this 785 is not human friendly/readable. One way to generate a Resource-ID 786 is by applying a mapping function to some other unique name (e.g., 787 user name or service name) for the resource. The Resource-ID is 788 used by the distributed database algorithm to determine the peer 789 or peers that are responsible for storing the data for the 790 overlay. In structured P2P networks, Resource-IDs are generally 791 fixed length and are formed by hashing the resource name. In 792 unstructured networks, resource names may be used directly as 793 Resource-IDs and may be variable lengths. 795 Connection Table: The set of nodes to which a node is directly 796 connected. This includes nodes with which Attach handshakes have 797 been done but which have not sent any Updates. 799 Routing Table: The set of peers which a node can use to route 800 overlay messages. In general, these peers will all be on the 801 connection table but not vice versa, because some peers will have 802 Attached but not sent updates. Peers may send messages directly 803 to peers that are in the connection table but may only route 804 messages to other peers through peers that are in the routing 805 table. 807 Destination List: A list of IDs through which a message is to be 808 routed. A single Node-ID is a trivial form of destination list. 810 Usage: A usage is an application that wishes to use the overlay for 811 some purpose. Each application wishing to use the overlay defines 812 a set of data kinds that it wishes to use. The SIP usage defines 813 the location data kind. 815 The term "maximum request lifetime" is the maximum time a request 816 will wait for a response; it defaults to 15 seconds. The term 817 "successor replacement hold-down time" is the amount of time to wait 818 before starting replication when a new successor is found; it 819 defaults to 30 seconds. 821 3. Overlay Management Overview 823 The most basic function of RELOAD is as a generic overlay network. 824 Nodes need to be able to join the overlay, form connections to other 825 nodes, and route messages through the overlay to nodes to which they 826 are not directly connected. This section provides an overview of the 827 mechanisms that perform these functions. 829 3.1. Security and Identification 831 Every node in the RELOAD overlay is identified by a Node-ID. The 832 Node-ID is used for three major purposes: 834 o To address the node itself. 835 o To determine its position in the overlay topology when the overlay 836 is structured. 837 o To determine the set of resources for which the node is 838 responsible. 840 Each node has a certificate [RFC5280] containing a Node-ID, which is 841 unique within an overlay instance. 843 The certificate serves multiple purposes: 845 o It entitles the user to store data at specific locations in the 846 Overlay Instance. Each data kind defines the specific rules for 847 determining which certificates can access each Resource-ID/Kind-ID 848 pair. For instance, some kinds might allow anyone to write at a 849 given location, whereas others might restrict writes to the owner 850 of a single certificate. 852 o It entitles the user to operate a node that has a Node-ID found in 853 the certificate. When the node forms a connection to another 854 peer, it uses this certificate so that a node connecting to it 855 knows it is connected to the correct node (technically: a (D)TLS 856 association with client authentication is formed.) In addition, 857 the node can sign messages, thus providing integrity and 858 authentication for messages which are sent from the node. 859 o It entitles the user to use the user name found in the 860 certificate. 862 If a user has more than one device, typically they would get one 863 certificate for each device. This allows each device to act as a 864 separate peer. 866 RELOAD supports multiple certificate issuance models. The first is 867 based on a central enrollment process which allocates a unique name 868 and Node-ID and puts them in a certificate for the user. All peers 869 in a particular Overlay Instance have the enrollment server as a 870 trust anchor and so can verify any other peer's certificate. 872 In some settings, a group of users want to set up an overlay network 873 but are not concerned about attack by other users in the network. 874 For instance, users on a LAN might want to set up a short term ad hoc 875 network without going to the trouble of setting up an enrollment 876 server. RELOAD supports the use of self-generated, self-signed 877 certificates. When self-signed certificates are used, the node also 878 generates its own Node-ID and username. The Node-ID is computed as a 879 digest of the public key, to prevent Node-ID theft; however this 880 model is still subject to a number of known attacks (most notably 881 Sybil attacks [Sybil]) and can only be safely used in closed networks 882 where users are mutually trusting. 884 The general principle here is that the security mechanisms (TLS and 885 message signatures) are always used, even if the certificates are 886 self-signed. This allows for a single set of code paths in the 887 systems with the only difference being whether certificate 888 verification is required to chain to a single root of trust. 890 3.1.1. Shared-Key Security 892 RELOAD also provides an admission control system based on shared 893 keys. In this model, the peers all share a single key which is used 894 to authenticate the peer-to-peer connections via TLS-PSK/TLS-SRP. 896 3.2. Clients 898 RELOAD defines a single protocol that is used both as the peer 899 protocol and as the client protocol for the overlay. This simplifies 900 implementation, particularly for devices that may act in either role, 901 and allows clients to inject messages directly into the overlay. 903 We use the term "peer" to identify a node in the overlay that routes 904 messages for nodes other than those to which it is directly 905 connected. Peers typically also have storage responsibilities. We 906 use the term "client" to refer to nodes that do not have routing or 907 storage responsibilities. When text applies to both peers and 908 clients, we will simply refer such devices as "nodes." 910 RELOAD's client support allows nodes that are not participating in 911 the overlay as peers to utilize the same implementation and to 912 benefit from the same security mechanisms as the peers. Clients 913 possess and use certificates that authorize the user to store data at 914 certain locations in the overlay. The Node-ID in the certificate is 915 used to identify the particular client as a member of the overlay and 916 to authenticate its messages. 918 In RELOAD, unlike some other designs, clients are not a first-class 919 concept. From the perspective of a peer, a client is simply a node 920 which has not yet sent any Updates or Joins. It might never do so 921 (if it's a client) or it might eventually do so (if it's just a node 922 that's taking a long time to join). The routing and storage rules 923 for RELOAD provide for correct behavior by peers regardless of 924 whether other nodes attached to them are clients or peers. Of 925 course, a client implementation must know that it intends to be a 926 client, but this localizes complexity only to that node. 928 For more discussion of the motivation for RELOAD's client support, 929 see Appendix B. 931 3.2.1. Client Routing 933 Clients may insert themselves in the overlay in two ways: 935 o Establish a connection to the peer responsible for the client's 936 Node-ID in the overlay. Then requests may be sent from/to the 937 client using its Node-ID in the same manner as if it were a peer, 938 because the responsible peer in the overlay will handle the final 939 step of routing to the client. This may require a TURN relay in 940 cases where NATs or firewalls prevent a client from forming a 941 direct connections with its responsible peer. Note that clients 942 that choose this option need to process Update messages from the 943 peer. Those updates can indicate that the peer no longer is 944 responsible for the Client's Node-ID. The client would then need 945 to form a connection to the appropriate peer. Failure to do so 946 will result in the client no longer receiving messages. 948 o Establish a connection with an arbitrary peer in the overlay 949 (perhaps based on network proximity or an inability to establish a 950 direct connection with the responsible peer). In this case, the 951 client will rely on RELOAD's Destination List feature to ensure 952 reachability. The client can initiate requests, and any node in 953 the overlay that knows the Destination List to its current 954 location can reach it, but the client is not directly reachable 955 using only its Node-ID. If the client is to receive incoming 956 requests from other members of the overlay, the Destination List 957 required to reach it must be learnable via other mechanisms, such 958 as being stored in the overlay by a usage. A client connected 959 this way using a certificate with only a single Node-ID MAY 960 proceed to use the connection without performing an Attach. A 961 client wishing to connect using this mechanism with a certificate 962 with multiple Node-IDs MUST perform an Attach after using Ping to 963 probe the Node-ID of the node to which it is connected. 965 3.2.2. Minimum Functionality Requirements for Clients 967 A node may act as a client simply because it does not have the 968 resources or even an implementation of the topology plugin required 969 to act as a peer in the overlay. In order to exchange RELOAD 970 messages with a peer, a client must meet a minimum level of 971 functionality. Such a client must: 973 o Implement RELOAD's connection-management operations that are used 974 to establish the connection with the peer. 975 o Implement RELOAD's data retrieval methods (with client 976 functionality). 977 o Be able to calculate Resource-IDs used by the overlay. 978 o Possess security credentials required by the overlay it is 979 implementing. 981 A client speaks the same protocol as the peers, knows how to 982 calculate Resource-IDs, and signs its requests in the same manner as 983 peers. While a client does not necessarily require a full 984 implementation of the overlay algorithm, calculating the Resource-ID 985 requires an implementation of the appropriate algorithm for the 986 overlay. 988 3.3. Routing 990 This section will discuss the requirements RELOAD's routing 991 capabilities were designed to meet, then describe the routing 992 features in the protocol, and then provide a brief overview of how 993 they are used. Appendix A discusses some alternative designs and the 994 tradeoffs that would be necessary to support them. 996 RELOAD's routing capabilities must meet the following requirements: 998 NAT Traversal: RELOAD must support establishing and using 999 connections between nodes separated by one or more NATs, including 1000 locating peers behind NATs for those overlays allowing/requiring 1001 it. 1002 Clients: RELOAD must support requests from and to clients that do 1003 not participate in overlay routing. 1004 Client promotion: RELOAD must support clients that become peers at a 1005 later point as determined by the overlay algorithm and deployment. 1006 Low state: RELOAD's routing algorithms must not require 1007 significant state to be stored on intermediate peers. 1008 Return routability in unstable topologies: At some points in 1009 times, different nodes may have inconsistent information about the 1010 connectivity of the routing graph. In all cases, the response to 1011 a request needs to delivered to the node that sent the request and 1012 not to some other node. 1014 RELOAD's routing provides three mechanisms designed to assist in 1015 meeting these needs: 1017 Destination Lists: While in principle it is possible to just 1018 inject a message into the overlay with a bare Node-ID as the 1019 destination, RELOAD provides a source routing capability in the 1020 form of "Destination Lists". A "Destination List provides a list 1021 of the nodes through which a message must flow. 1022 Via Lists: In order to allow responses to follow the same path as 1023 requests, each message also contains a "Via List", which is added 1024 to by each node a message traverses. This via list can then be 1025 inverted and used as a destination list for the response. 1026 RouteQuery: The RouteQuery method allows a node to query a peer 1027 for the next hop it will use to route a message. This method is 1028 useful for diagnostics and for iterative routing. 1030 The basic routing mechanism used by RELOAD is Symmetric Recursive. 1031 We will first describe symmetric recursive routing and then discuss 1032 its advantages in terms of the requirements discussed above. 1034 Symmetric recursive routing requires that a request message follow a 1035 path through the overlay to the destination: each peer forwards the 1036 message closer to its destination. The return path of the response 1037 is then the same path followed in reverse. For example, a message 1038 following a route from A to Z through B and X: 1040 A B X Z 1041 ------------------------------- 1043 ----------> 1044 Dest=Z 1045 ----------> 1046 Via=A 1047 Dest=Z 1048 ----------> 1049 Via=A, B 1050 Dest=Z 1052 <---------- 1053 Dest=X, B, A 1054 <---------- 1055 Dest=B, A 1056 <---------- 1057 Dest=A 1059 Note that the preceding Figure does not indicate whether A is a 1060 client or peer: A forwards its request to B and the response is 1061 returned to A in the same manner regardless of A's role in the 1062 overlay. 1064 This figure shows use of full via-lists by intermediate peers B and 1065 X. However, if B and/or X are willing to store state, then they may 1066 elect to truncate the lists, save that information internally (keyed 1067 by the transaction id), and return the response message along the 1068 path from which it was received when the response is received. This 1069 option requires greater state to be stored on intermediate peers but 1070 saves a small amount of bandwidth and reduces the need for modifying 1071 the message en route. Selection of this mode of operation is a 1072 choice for the individual peer; the techniques are interoperable even 1073 on a single message. The figure below shows B using full via lists 1074 but X truncating them to X1 and saving the state internally. 1076 A B X Z 1077 ------------------------------- 1079 ----------> 1080 Dest=Z 1081 ----------> 1082 Via=A 1083 Dest=Z 1084 ----------> 1085 Dest=Z, X1 1087 <---------- 1088 Dest=X,X1 1089 <---------- 1090 Dest=B, A 1091 <---------- 1092 Dest=A 1094 RELOAD also supports a basic Iterative routing mode (where the 1095 intermediate peers merely return a response indicating the next hop, 1096 but do not actually forward the message to that next hop themselves). 1097 Iterative routing is implemented using the RouteQuery method, which 1098 requests this behavior. Note that iterative routing is selected only 1099 by the initiating node. 1101 3.4. Connectivity Management 1103 In order to provide efficient routing, a peer needs to maintain a set 1104 of direct connections to other peers in the Overlay Instance. Due to 1105 the presence of NATs, these connections often cannot be formed 1106 directly. Instead, we use the Attach request to establish a 1107 connection. Attach uses ICE [RFC5245] to establish the connection. 1108 It is assumed that the reader is familiar with ICE. 1110 Say that peer A wishes to form a direct connection to peer B. It 1111 gathers ICE candidates and packages them up in an Attach request 1112 which it sends to B through usual overlay routing procedures. B does 1113 its own candidate gathering and sends back a response with its 1114 candidates. A and B then do ICE connectivity checks on the candidate 1115 pairs. The result is a connection between A and B. At this point, A 1116 and B can add each other to their routing tables and send messages 1117 directly between themselves without going through other overlay 1118 peers. 1120 There are two cases where Attach is not used. The first is when a 1121 peer is joining the overlay and is not connected to any peers. In 1122 order to support this case, some small number of "bootstrap nodes" 1123 typically need to be publicly accessible so that new peers can 1124 directly connect to them. Section 10 contains more detail on this. 1125 The second case is when a client node connects to a node at an 1126 arbitrary IP address, rather than to its responsible peer, as 1127 described in the second bullet point of Section 3.2.1. 1129 In general, a peer needs to maintain connections to all of the peers 1130 near it in the Overlay Instance and to enough other peers to have 1131 efficient routing (the details depend on the specific overlay). If a 1132 peer cannot form a connection to some other peer, this isn't 1133 necessarily a disaster; overlays can route correctly even without 1134 fully connected links. However, a peer should try to maintain the 1135 specified link set and if it detects that it has fewer direct 1136 connections, should form more as required. This also implies that 1137 peers need to periodically verify that the connected peers are still 1138 alive and if not try to reform the connection or form an alternate 1139 one. 1141 3.5. Overlay Algorithm Support 1143 The Topology Plugin allows RELOAD to support a variety of overlay 1144 algorithms. This specification defines a DHT based on Chord [Chord], 1145 which is mandatory to implement, but the base RELOAD protocol is 1146 designed to support a variety of overlay algorithms. 1148 3.5.1. Support for Pluggable Overlay Algorithms 1150 RELOAD defines three methods for overlay maintenance: Join, Update, 1151 and Leave. However, the contents of those messages, when they are 1152 sent, and their precise semantics are specified by the actual overlay 1153 algorithm; RELOAD merely provides a framework of commonly-needed 1154 methods that provides uniformity of notation (and ease of debugging) 1155 for a variety of overlay algorithms. 1157 3.5.2. Joining, Leaving, and Maintenance Overview 1159 When a new peer wishes to join the Overlay Instance, it must have a 1160 Node-ID that it is allowed to use and a set of credentials which 1161 match that Node-ID. When an enrollment server is used that Node-ID 1162 will be in the certificate the node received from the enrollment 1163 server. The details of the joining procedure are defined by the 1164 overlay algorithm, but the general steps for joining an Overlay 1165 Instance are: 1167 o Forming connections to some other peers. 1168 o Acquiring the data values this peer is responsible for storing. 1169 o Informing the other peers which were previously responsible for 1170 that data that this peer has taken over responsibility. 1172 The first thing the peer needs to do is to form a connection to some 1173 "bootstrap node". Because this is the first connection the peer 1174 makes, these nodes must have public IP addresses so that they can be 1175 connected to directly. Once a peer has connected to one or more 1176 bootstrap nodes, it can form connections in the usual way by routing 1177 Attach messages through the overlay to other nodes. Once a peer has 1178 connected to the overlay for the first time, it can cache the set of 1179 nodes it has connected to with public IP addresses for use as future 1180 bootstrap nodes. 1182 Once a peer has connected to a bootstrap node, it then needs to take 1183 up its appropriate place in the overlay. This requires two major 1184 operations: 1186 o Forming connections to other peers in the overlay to populate its 1187 Routing Table. 1188 o Getting a copy of the data it is now responsible for storing and 1189 assuming responsibility for that data. 1191 The second operation is performed by contacting the Admitting Peer 1192 (AP), the node which is currently responsible for that section of the 1193 overlay. 1195 The details of this operation depend mostly on the overlay algorithm 1196 involved, but a typical case would be: 1198 1. JP (Joining Peer) sends a Join request to AP (Admitting Peer) 1199 announcing its intention to join. 1200 2. AP sends a Join response. 1201 3. AP does a sequence of Stores to JP to give it the data it will 1202 need. 1203 4. AP does Updates to JP and to other peers to tell it about its own 1204 routing table. At this point, both JP and AP consider JP 1205 responsible for some section of the Overlay Instance. 1206 5. JP makes its own connections to the appropriate peers in the 1207 Overlay Instance. 1209 After this process is completed, JP is a full member of the Overlay 1210 Instance and can process Store/Fetch requests. 1212 Note that the first node is a special case. When ordinary nodes 1213 cannot form connections to the bootstrap nodes, then they are not 1214 part of the overlay. However, the first node in the overlay can 1215 obviously not connect to other nodes. In order to support this case, 1216 potential first nodes (which must also serve as bootstrap nodes 1217 initially) must somehow be instructed (perhaps by configuration 1218 settings) that they are the entire overlay, rather than not part of 1219 it. 1221 Note that clients do not perform either of these operations. 1223 3.6. First-Time Setup 1225 Previous sections addressed how RELOAD works once a node has 1226 connected. This section provides an overview of how users get 1227 connected to the overlay for the first time. RELOAD is designed so 1228 that users can start with the name of the overlay they wish to join 1229 and perhaps a username and password, and leverage that into having a 1230 working peer with minimal user intervention. This helps avoid the 1231 problems that have been experienced with conventional SIP clients 1232 where users are required to manually configure a large number of 1233 settings. 1235 3.6.1. Initial Configuration 1237 In the first phase of the process, the user starts out with the name 1238 of the overlay and uses this to download an initial set of overlay 1239 configuration parameters. The node does a DNS SRV lookup on the 1240 overlay name to get the address of a configuration server. It can 1241 then connect to this server with HTTPS to download a configuration 1242 document which contains the basic overlay configuration parameters as 1243 well as a set of bootstrap nodes which can be used to join the 1244 overlay. 1246 If a node already has the valid configuration document that it 1247 received by some out of band method, this step can be skipped. 1249 3.6.2. Enrollment 1251 If the overlay is using centralized enrollment, then a user needs to 1252 acquire a certificate before joining the overlay. The certificate 1253 attests both to the user's name within the overlay and to the Node- 1254 IDs which they are permitted to operate. In that case, the 1255 configuration document will contain the address of an enrollment 1256 server which can be used to obtain such a certificate. The 1257 enrollment server may (and probably will) require some sort of 1258 username and password before issuing the certificate. The enrollment 1259 server's ability to restrict attackers' access to certificates in the 1260 overlay is one of the cornerstones of RELOAD's security. 1262 4. Application Support Overview 1264 RELOAD is not intended to be used alone, but rather as a substrate 1265 for other applications. These applications can use RELOAD for a 1266 variety of purposes: 1268 o To store data in the overlay and retrieve data stored by other 1269 nodes. 1270 o As a discovery mechanism for services such as TURN. 1271 o To form direct connections which can be used to transmit 1272 application-level messages without using the overlay. 1274 This section provides an overview of these services. 1276 4.1. Data Storage 1278 RELOAD provides operations to Store and Fetch data. Each location in 1279 the Overlay Instance is referenced by a Resource-ID. However, each 1280 location may contain data elements corresponding to multiple kinds 1281 (e.g., certificate, SIP registration). Similarly, there may be 1282 multiple elements of a given kind, as shown below: 1284 +--------------------------------+ 1285 | Resource-ID | 1286 | | 1287 | +------------+ +------------+ | 1288 | | Kind 1 | | Kind 2 | | 1289 | | | | | | 1290 | | +--------+ | | +--------+ | | 1291 | | | Value | | | | Value | | | 1292 | | +--------+ | | +--------+ | | 1293 | | | | | | 1294 | | +--------+ | | +--------+ | | 1295 | | | Value | | | | Value | | | 1296 | | +--------+ | | +--------+ | | 1297 | | | +------------+ | 1298 | | +--------+ | | 1299 | | | Value | | | 1300 | | +--------+ | | 1301 | +------------+ | 1302 +--------------------------------+ 1304 Each kind is identified by a Kind-ID, which is a code point either 1305 assigned by IANA or allocated out of a private range. As part of the 1306 kind definition, protocol designers may define constraints, such as 1307 limits on size, on the values which may be stored. For many kinds, 1308 the set may be restricted to a single value; some sets may be allowed 1309 to contain multiple identical items while others may only have unique 1310 items. Note that a kind may be employed by multiple usages and new 1311 usages are encouraged to use previously defined kinds where possible. 1312 We define the following data models in this document, though other 1313 usages can define their own structures: 1315 single value: There can be at most one item in the set and any value 1316 overwrites the previous item. 1318 array: Many values can be stored and addressed by a numeric index. 1320 dictionary: The values stored are indexed by a key. Often this key 1321 is one of the values from the certificate of the peer sending the 1322 Store request. 1324 In order to protect stored data from tampering, by other nodes, each 1325 stored value is digitally signed by the node which created it. When 1326 a value is retrieved, the digital signature can be verified to detect 1327 tampering. 1329 4.1.1. Storage Permissions 1331 A major issue in peer-to-peer storage networks is minimizing the 1332 burden of becoming a peer, and in particular minimizing the amount of 1333 data which any peer is required to store for other nodes. RELOAD 1334 addresses this issue by only allowing any given node to store data at 1335 a small number of locations in the overlay, with those locations 1336 being determined by the node's certificate. When a peer uses a Store 1337 request to place data at a location authorized by its certificate, it 1338 signs that data with the private key that corresponds to its 1339 certificate. Then the peer responsible for storing the data is able 1340 to verify that the peer issuing the request is authorized to make 1341 that request. Each data kind defines the exact rules for determining 1342 what certificate is appropriate. 1344 The most natural rule is that a certificate authorizes a user to 1345 store data keyed with their user name X. This rule is used for all 1346 the kinds defined in this specification. Thus, only a user with a 1347 certificate for "alice@example.org" could write to that location in 1348 the overlay. However, other usages can define any rules they choose, 1349 including publicly writable values. 1351 The digital signature over the data serves two purposes. First, it 1352 allows the peer responsible for storing the data to verify that this 1353 Store is authorized. Second, it provides integrity for the data. 1354 The signature is saved along with the data value (or values) so that 1355 any reader can verify the integrity of the data. Of course, the 1356 responsible peer can "lose" the value but it cannot undetectably 1357 modify it. 1359 The size requirements of the data being stored in the overlay are 1360 variable. For instance, a SIP AOR and voicemail differ widely in the 1361 storage size. RELOAD leaves it to the Usage and overlay 1362 configuration to limit size imbalance of various kinds. 1364 4.1.2. Replication 1366 Replication in P2P overlays can be used to provide: 1368 persistence: if the responsible peer crashes and/or if the storing 1369 peer leaves the overlay 1370 security: to guard against DoS attacks by the responsible peer or 1371 routing attacks to that responsible peer 1372 load balancing: to balance the load of queries for popular 1373 resources. 1375 A variety of schemes are used in P2P overlays to achieve some of 1376 these goals. Common techniques include replicating on neighbors of 1377 the responsible peer, randomly locating replicas around the overlay, 1378 or replicating along the path to the responsible peer. 1380 The core RELOAD specification does not specify a particular 1381 replication strategy. Instead, the first level of replication 1382 strategies are determined by the overlay algorithm, which can base 1383 the replication strategy on its particular topology. For example, 1384 Chord places replicas on successor peers, which will take over 1385 responsibility should the responsible peer fail [Chord]. 1387 If additional replication is needed, for example if data persistence 1388 is particularly important for a particular usage, then that usage may 1389 specify additional replication, such as implementing random 1390 replications by inserting a different well known constant into the 1391 Resource Name used to store each replicated copy of the resource. 1392 Such replication strategies can be added independent of the 1393 underlying algorithm, and their usage can be determined based on the 1394 needs of the particular usage. 1396 4.2. Usages 1398 By itself, the distributed storage layer just provides infrastructure 1399 on which applications are built. In order to do anything useful, a 1400 usage must be defined. Each Usage needs to specify several things: 1402 o Registers Kind-ID code points for any kinds that the Usage 1403 defines. 1404 o Defines the data structure for each of the kinds. 1405 o Defines access control rules for each of the kinds. 1406 o Defines how the Resource Name is formed that is hashed to form the 1407 Resource-ID where each kind is stored. 1409 o Describes how values will be merged after a network partition. 1410 Unless otherwise specified, the default merging rule is to act as 1411 if all the values that need to be merged were stored and as if the 1412 order they were stored in corresponds to the stored time values 1413 associated with (and carried in) their values. Because the stored 1414 time values are those associated with the peer which did the 1415 writing, clock skew is generally not an issue. If two nodes are 1416 on different partitions, write to the same location, and have 1417 clock skew, this can create merge conflicts. However because 1418 RELOAD deliberately segregates storage so that data from different 1419 users and peers is stored in different locations, and a single 1420 peer will typically only be in a single network partition, this 1421 case will generally not arise. 1423 The kinds defined by a usage may also be applied to other usages. 1424 However, a need for different parameters, such as different size 1425 limits, would imply the need to create a new kind. 1427 4.3. Service Discovery 1429 RELOAD does not currently define a generic service discovery 1430 algorithm as part of the base protocol, although a simplistic TURN- 1431 specific discovery mechanism is provided. A variety of service 1432 discovery algorithms can be implemented as extensions to the base 1433 protocol, such as the service discovery algorithm ReDIR 1434 [opendht-sigcomm05] or [I-D.ietf-p2psip-service-discovery]. 1436 4.4. Application Connectivity 1438 There is no requirement that a RELOAD usage must use RELOAD's 1439 primitives for establishing its own communication if it already 1440 possesses its own means of establishing connections. For example, 1441 one could design a RELOAD-based resource discovery protocol which 1442 used HTTP to retrieve the actual data. 1444 For more common situations, however, it is the overlay itself - 1445 rather than an external authority such as DNS - which is used to 1446 establish a connection. RELOAD provides connectivity to applications 1447 using the AppAttach method. For example, if a P2PSIP node wishes to 1448 establish a SIP dialog with another P2PSIP node, it will use 1449 AppAttach to establish a direct connection with the other node. This 1450 new connection is separate from the peer protocol connection. It is 1451 a dedicated UDP or TCP flow used only for the SIP dialog. 1453 5. Overlay Management Protocol 1455 This section defines the basic protocols used to create, maintain, 1456 and use the RELOAD overlay network. We start by defining the basic 1457 concept of how message destinations are interpreted when routing 1458 messages. We then describe the symmetric recursive routing model, 1459 which is RELOAD's default routing algorithm. We then define the 1460 message structure and then finally define the messages used to join 1461 and maintain the overlay. 1463 5.1. Message Receipt and Forwarding 1465 When a peer receives a message, it first examines the overlay, 1466 version, and other header fields to determine whether the message is 1467 one it can process. If any of these are incorrect (e.g., the message 1468 is for an overlay in which the peer does not participate) it is an 1469 error. The peer SHOULD generate an appropriate error but local 1470 policy can override this and cause the messages to be silently 1471 dropped. 1473 Once the peer has determined that the message is correctly formatted, 1474 it examines the first entry on the destination list. There are three 1475 possible cases here: 1477 o The first entry on the destination list is an ID for which the 1478 peer is responsible. A peer is always responsible for the 1479 wildcard Node-ID. Handling of this case is described in 1480 Section 5.1.1. 1481 o The first entry on the destination list is an ID for which another 1482 peer is responsible. Handling of this case is described in 1483 Section 5.1.2. 1484 o The first entry on the destination list is a private ID that is 1485 being used for destination list compression. Handling of this 1486 case is described in Section 5.1.3. Note that private IDs can be 1487 distinguished from Node-IDs and Resource-IDs on the wire as 1488 described in Section 5.3.2.2). 1490 These cases are handled as discussed below. 1492 5.1.1. Responsible ID 1494 If the first entry on the destination list is an ID for which the 1495 peer is responsible, there are several sub-cases to consider. 1497 o If the entry is a Resource-ID, then it MUST be the only entry on 1498 the destination list. If there are other entries, the message 1499 MUST be silently dropped. Otherwise, the message is destined for 1500 this node and it passes it up to the upper layers. 1501 o If the entry is a Node-ID which equals this node's Node-ID, then 1502 the message is destined for this node. If this is the only entry 1503 on the destination list, the message is destined for this node and 1504 is passed up to the upper layers. Otherwise the entry is removed 1505 from the destination list and the message is passed to the Message 1506 Transport. If the message is a response and there is state for 1507 the transaction ID, the state is reinserted into the destination 1508 list before the message is further processed. 1509 o If the entry is the wildcard Node-ID, the message is destined for 1510 this node and it passes it up to the upper layers. 1511 o If the entry is a Node-ID which is not equal to this node, then 1512 the node MUST drop the message silently unless the Node-ID 1513 corresponds to a node which is directly connected to this node 1514 (i.e., a client). In that case, it MUST forward the message to 1515 the destination node as described in the next section. 1517 Note that this implies that in order to address a message to "the 1518 peer that controls region X", a sender sends to Resource-ID X, not 1519 Node-ID X. 1521 5.1.2. Other ID 1523 If neither of the other three cases applies, then the peer MUST 1524 forward the message towards the first entry on the destination list. 1525 This means that it MUST select one of the peers to which it is 1526 connected and which is likely to be responsible for the first entry 1527 on the destination list. If the first entry on the destination list 1528 is in the peer's connection table, then it SHOULD forward the message 1529 to that peer directly. Otherwise, the peer consults the routing 1530 table to forward the message. 1532 Any intermediate peer which forwards a RELOAD request MUST arrange 1533 that if it receives a response to that message the response can be 1534 routed back through the set of nodes through which the request 1535 passed. This may be arranged in one of two ways: 1537 o The peer MAY add an entry to the via list in the forwarding header 1538 that will enable it to determine the correct node. 1539 o The peer MAY keep per-transaction state which will allow it to 1540 determine the correct node. 1542 As an example of the first strategy, if node D receives a message 1543 from node C with via list (A, B), then D would forward to the next 1544 node (E) with via list (A, B, C). Now, if E wants to respond to the 1545 message, it reverses the via list to produce the destination list, 1546 resulting in (D, C, B, A). When D forwards the response to C, the 1547 destination list will contain (C, B, A). 1549 As an example of the second strategy, if node D receives a message 1550 from node C with transaction ID X and via list (A, B), it could store 1551 (X, C) in its state database and forward the message with the via 1552 list unchanged. When D receives the response, it consults its state 1553 database for transaction id X, determines that the request came from 1554 C, and forwards the response to C. 1556 Intermediate peers which modify the via list are not required to 1557 simply add entries. The only requirement is that the peer be able to 1558 reconstruct the correct destination list on the return route. RELOAD 1559 provides explicit support for this functionality in the form of 1560 private IDs, which can replace any number of via list entries. For 1561 instance, in the above example, Node D might send E a via list 1562 containing only the private ID (I). E would then use the destination 1563 list (D, I) to send its return message. When D processes this 1564 destination list, it would detect that I is a private ID, recover the 1565 via list (A, B, C), and reverse that to produce the correct 1566 destination list (C, B, A) before sending it to C. This feature is 1567 called List Compression. It MAY either be a compressed version of 1568 the original via list or an index into a state database containing 1569 the original via list. 1571 No matter what mechanism for storing via list state is used, if an 1572 intermediate peer exits the overlay, then on the return trip the 1573 message cannot be forwarded and will be dropped. The ordinary 1574 timeout and retransmission mechanisms provide stability over this 1575 type of failure. 1577 Note that if an intermediate peer retains per-transaction state 1578 instead of modifying the via list, it needs some mechanism for timing 1579 out that state, otherwise its state database will grow without bound. 1580 Whatever algorithm is used, unless a FORWARD_CRITICAL forwarding 1581 option or overlay configuration option explicitly indicates this 1582 state is not needed, the state MUST be maintained for at least the 1583 value of the overlay reliability timer (3 seconds) and MAY be kept 1584 longer. Future extension, such as [I-D.jiang-p2psip-relay], may 1585 define mechanisms for determining when this state does not need to be 1586 retained. 1588 None of the above mechanisms are required for responses, since there 1589 is no need to ensure that subsequent requests follow the same path. 1591 To be precise on the responsibility of the intermediate node, suppose 1592 that an intermediate node, A, receives a message from node B with via 1593 list X-Y-Z. Node A MUST implement an algorithm that ensures that A 1594 returns a response to this request to node B with the destination 1595 list B-Z-Y-X, provided that the node to which A forwards the request 1596 follows the same contract. Node A normally learns the Node-ID B is 1597 using via an Attach, but a node using a certificate with a single 1598 Node-ID MAY elect to not send an Attach (see Section 3.2.1 bullet 2). 1599 If a node with a certificate with multiple Node-IDs attempts to route 1600 a message other than a Ping or Attach through a node without 1601 performing an Attach, the receiving node MUST reject the request with 1602 an Error_Forbidden error. The node MUST implement support for 1603 returning responses to a Ping or Attach request made by a joining 1604 node Attaching to its responsible peer. 1606 5.1.3. Private ID 1608 If the first entry in the destination list is a private id (e.g., a 1609 compressed via list), the peer MUST replace that entry with the 1610 original via list that it replaced and then re-examine the 1611 destination list to determine which of the three cases in Section 5.1 1612 now applies. 1614 5.2. Symmetric Recursive Routing 1616 This Section defines RELOAD's symmetric recursive routing algorithm, 1617 which is the default algorithm used by nodes to route messages 1618 through the overlay. All implementations MUST implement this routing 1619 algorithm. An overlay may be configured to use alternative routing 1620 algorithms, and alternative routing algorithms may be selected on a 1621 per-message basis. 1623 5.2.1. Request Origination 1625 In order to originate a message to a given Node-ID or Resource-ID, a 1626 node constructs an appropriate destination list. The simplest such 1627 destination list is a single entry containing the Node-ID or 1628 Resource-ID. The resulting message will use the normal overlay 1629 routing mechanisms to forward the message to that destination. The 1630 node can also construct a more complicated destination list for 1631 source routing. 1633 Once the message is constructed, the node sends the message to some 1634 adjacent peer. If the first entry on the destination list is 1635 directly connected, then the message MUST be routed down that 1636 connection. Otherwise, the topology plugin MUST be consulted to 1637 determine the appropriate next hop. 1639 Parallel searches for the resource are a common solution to improve 1640 reliability in the face of churn or of subversive peers. Parallel 1641 searches for usage-specified replicas are managed by the usage layer. 1642 However, a single request can also be routed through multiple 1643 adjacent peers, even when known to be sub-optimal, to improve 1644 reliability [vulnerabilities-acsac04]. Such parallel searches MAY be 1645 specified by the topology plugin. 1647 Because messages may be lost in transit through the overlay, RELOAD 1648 incorporates an end-to-end reliability mechanism. When an 1649 originating node transmits a request it MUST set a 3 second timer. 1650 If a response has not been received when the timer fires, the request 1651 is retransmitted with the same transaction identifier. The request 1652 MAY be retransmitted up to 4 times (for a total of 5 messages). 1653 After the timer for the fifth transmission fires, the message SHALL 1654 be considered to have failed. Note that this retransmission 1655 procedure is not followed by intermediate nodes. They follow the 1656 hop-by-hop reliability procedure described in Section 5.6.3. 1658 The above algorithm can result in multiple requests being delivered 1659 to a node. Receiving nodes MUST generate semantically equivalent 1660 responses to retransmissions of the same request (this can be 1661 determined by transaction id) if the request is received within the 1662 maximum request lifetime (15 seconds). For some requests (e.g., 1663 Fetch) this can be accomplished merely by processing the request 1664 again. For other requests, (e.g., Store) it may be necessary to 1665 maintain state for the duration of the request lifetime. 1667 5.2.2. Response Origination 1669 When a peer sends a response to a request using this routing 1670 algorithm, it MUST construct the destination list by reversing the 1671 order of the entries on the via list. This has the result that the 1672 response traverses the same peers as the request traversed, except in 1673 reverse order (symmetric routing). 1675 5.3. Message Structure 1677 RELOAD is a message-oriented request/response protocol. The messages 1678 are encoded using binary fields. All integers are represented in 1679 network byte order. The general philosophy behind the design was to 1680 use Type, Length, Value fields to allow for extensibility. However, 1681 for the parts of a structure that were required in all messages, we 1682 just define these in a fixed position, as adding a type and length 1683 for them is unnecessary and would simply increase bandwidth and 1684 introduces new potential for interoperability issues. 1686 Each message has three parts, concatenated as shown below: 1688 +-------------------------+ 1689 | Forwarding Header | 1690 +-------------------------+ 1691 | Message Contents | 1692 +-------------------------+ 1693 | Security Block | 1694 +-------------------------+ 1696 The contents of these parts are as follows: 1698 Forwarding Header: Each message has a generic header which is used 1699 to forward the message between peers and to its final destination. 1700 This header is the only information that an intermediate peer 1701 (i.e., one that is not the target of a message) needs to examine. 1703 Message Contents: The message being delivered between the peers. 1704 From the perspective of the forwarding layer, the contents are 1705 opaque, however, they are interpreted by the higher layers. 1707 Security Block: A security block containing certificates and a 1708 digital signature over the "Message Contents" section. Note that 1709 this signature can be computed without parsing the message 1710 contents. All messages MUST be signed by their originator. 1712 The following sections describe the format of each part of the 1713 message. 1715 5.3.1. Presentation Language 1717 The structures defined in this document are defined using a C-like 1718 syntax based on the presentation language used to define 1719 TLS[RFC5246]. Advantages of this style include: 1721 o It familiar enough looking that most readers can grasp it quickly. 1722 o The ability to define nested structures allows a separation 1723 between high-level and low-level message structures. 1724 o It has a straightforward wire encoding that allows quick 1725 implementation, but the structures can be comprehended without 1726 knowing the encoding. 1727 o The ability to mechanically compile encoders and decoders. 1729 Several idiosyncrasies of this language are worth noting. 1731 o All lengths are denoted in bytes, not objects. 1732 o Variable length values are denoted like arrays with angle 1733 brackets. 1734 o "select" is used to indicate variant structures. 1736 For instance, "uint16 array<0..2^8-2>;" represents up to 254 bytes 1737 which corresponds to up to 127 values of two bytes (16 bits) each. 1739 5.3.1.1. Common Definitions 1741 The following definitions are used throughout RELOAD and so are 1742 defined here. They also provide a convenient introduction to how to 1743 read the presentation language. 1745 An enum represents an enumerated type. The values associated with 1746 each possibility are represented in parentheses and the maximum value 1747 is represented as a nameless value, for purposes of describing the 1748 width of the containing integral type. For instance, Boolean 1749 represents a true or false: 1751 enum { false (0), true(1), (255)} Boolean; 1753 A boolean value is either a 1 or a 0. The max value of 255 indicates 1754 this is represented as a single byte on the wire. 1756 The NodeId, shown below, represents a single Node-ID. 1758 typedef opaque NodeId[NodeIdLength]; 1760 A NodeId is a fixed-length structure represented as a series of 1761 bytes, with the most significant byte first. The length is set on a 1762 per-overlay basis within the range of 16-20 bytes (128 to 160 bits). 1763 (See Section 10.1 for how NodeIdLength is set.) Note: the use of 1764 "typedef" here is an extension to the TLS language, but its meaning 1765 should be relatively obvious. Note the [ size ] syntax defines a 1766 fixed length element that does not include the length of the element 1767 in the on the wire encoding. 1769 A ResourceId, shown below, represents a single Resource-ID. 1771 typedef opaque ResourceId<0..2^8-1>; 1773 Like a NodeId, a ResourceId is an opaque string of bytes, but unlike 1774 NodeIds, ResourceIds are variable length, up to 254 bytes (2040 bits) 1775 in length. On the wire, each ResourceId is preceded by a single 1776 length byte (allowing lengths up to 255). Thus, the 3-byte value 1777 "FOO" would be encoded as: 03 46 4f 4f. Note the < range > syntax 1778 defines a variable length element that does include the length of the 1779 element in the on the wire encoding. The number of bytes to encode 1780 the length on the wire is derived by range; i.e., it is the minimum 1781 number of bytes which can encode the largest range value. 1783 A more complicated example is IpAddressPort, which represents a 1784 network address and can be used to carry either an IPv6 or IPv4 1785 address: 1787 enum {reservedAddr(0), ipv4_address (1), ipv6_address (2), 1788 (255)} AddressType; 1790 struct { 1791 uint32 addr; 1792 uint16 port; 1793 } IPv4AddrPort; 1795 struct { 1796 uint128 addr; 1797 uint16 port; 1798 } IPv6AddrPort; 1800 struct { 1801 AddressType type; 1802 uint8 length; 1804 select (type) { 1805 case ipv4_address: 1806 IPv4AddrPort v4addr_port; 1808 case ipv6_address: 1809 IPv6AddrPort v6addr_port; 1811 /* This structure can be extended */ 1812 }; 1813 } IpAddressPort; 1815 The first two fields in the structure are the same no matter what 1816 kind of address is being represented: 1818 type: the type of address (v4 or v6). 1819 length: the length of the rest of the structure. 1821 By having the type and the length appear at the beginning of the 1822 structure regardless of the kind of address being represented, an 1823 implementation which does not understand new address type X can still 1824 parse the IpAddressPort field and then discard it if it is not 1825 needed. 1827 The rest of the IpAddressPort structure is either an IPv4AddrPort or 1828 an IPv6AddrPort. Both of these simply consist of an address 1829 represented as an integer and a 16-bit port. As an example, here is 1830 the wire representation of the IPv4 address "192.0.2.1" with port 1831 "6100". 1833 01 ; type = IPv4 1834 06 ; length = 6 1835 c0 00 02 01 ; address = 192.0.2.1 1836 17 d4 ; port = 6100 1838 Unless a given structure that uses a select explicitly allows for 1839 unknown types in the select, any unknown type SHOULD be treated as an 1840 parsing error and the whole message discarded with no response. 1842 5.3.2. Forwarding Header 1844 The forwarding header is defined as a ForwardingHeader structure, as 1845 shown below. 1847 struct { 1848 uint32 relo_token; 1849 uint32 overlay; 1850 uint16 configuration_sequence; 1851 uint8 version; 1852 uint8 ttl; 1853 uint32 fragment; 1854 uint32 length; 1855 uint64 transaction_id; 1856 uint32 max_response_length; 1857 uint16 via_list_length; 1858 uint16 destination_list_length; 1859 uint16 options_length; 1860 Destination via_list[via_list_length]; 1861 Destination destination_list 1862 [destination_list_length]; 1863 ForwardingOptions options[options_length]; 1864 } ForwardingHeader; 1866 The contents of the structure are: 1868 relo_token: The first four bytes identify this message as a RELOAD 1869 message. This field MUST contain the value 0xd2454c4f (the string 1870 'RELO' with the high bit of the first byte set). 1872 overlay: The 32 bit checksum/hash of the overlay being used. The 1873 variable length string representing the overlay name is hashed 1874 with SHA-1 [RFC3174] and the low order 32 bits are used. The 1875 purpose of this field is to allow nodes to participate in multiple 1876 overlays and to detect accidental misconfiguration. This is not a 1877 security critical function. 1879 configuration_sequence: The sequence number of the configuration 1880 file. 1882 version: The version of the RELOAD protocol being used. This is a 1883 fixed point integer between 0.1 and 25.4. This document describes 1884 version 0.1, with a value of 0x01. [[ Note to RFC Editor: Please 1885 update this to version 1.0 with value of 0x0a and remove this 1886 note. ]] 1888 ttl: An 8 bit field indicating the number of iterations, or hops, a 1889 message can experience before it is discarded. The TTL value MUST 1890 be decremented by one at every hop along the route the message 1891 traverses. If the TTL is 0, the message MUST NOT be propagated 1892 further and MUST be discarded, and a "Error_TTL_Exceeded" error 1893 should be generated. The initial value of the TTL SHOULD be 100 1894 unless defined otherwise by the overlay configuration. 1896 fragment: This field is used to handle fragmentation. The high 1897 order two bits are used to indicate the fragmentation status: If 1898 the high bit (0x80000000) is set, it indicates that the message is 1899 a fragment. If the next bit (0x40000000) is set, it indicates 1900 that this is the last fragment. The next six bits (0x20000000 to 1901 0x01000000) are reserved and SHOULD be set to zero. The remainder 1902 of the field is used to indicate the fragment offset; see 1903 Section 5.7 1905 length: The count in bytes of the size of the message, including the 1906 header. 1908 transaction_id: A unique 64 bit number that identifies this 1909 transaction and also allows receivers to disambiguate transactions 1910 which are otherwise identical. In order to provide a high 1911 probability that transaction IDs are unique, they MUST be randomly 1912 generated. Responses use the same Transaction ID as the request 1913 they correspond to. Transaction IDs are also used for fragment 1914 reassembly. 1916 max_response_length: The maximum size in bytes of a response. Used 1917 by requesting nodes to avoid receiving (unexpected) very large 1918 responses. If this value is non-zero, responding peers MUST check 1919 that any response would not exceed it and if so generate an 1920 Error_Response_Too_Large value. This value SHOULD be set to zero 1921 for responses. 1923 via_list_length: The length of the via list in bytes. Note that in 1924 this field and the following two length fields we depart from the 1925 usual variable-length convention of having the length immediately 1926 precede the value in order to make it easier for hardware decoding 1927 engines to quickly determine the length of the header. 1929 destination_list_length: The length of the destination list in 1930 bytes. 1932 options_length: The length of the header options in bytes. 1934 via_list: The via_list contains the sequence of destinations through 1935 which the message has passed. The via_list starts out empty and 1936 grows as the message traverses each peer. 1938 destination_list: The destination_list contains a sequence of 1939 destinations which the message should pass through. The 1940 destination list is constructed by the message originator. The 1941 first element in the destination list is where the message goes 1942 next. The list shrinks as the message traverses each listed peer. 1944 options: Contains a series of ForwardingOptions entries. See 1945 Section 5.3.2.3. 1947 5.3.2.1. Processing Configuration Sequence Numbers 1949 In order to be part of the overlay, a node MUST have a copy of the 1950 overlay configuration document. In order to allow for configuration 1951 document changes, each version of the configuration document has a 1952 sequence number which is monotonically increasing mod 65536. Because 1953 the sequence number may in principle wrap, greater than or less than 1954 are interpreted by modulo arithmetic as in TCP. 1956 When a destination node receives a request, it MUST check that the 1957 configuration_sequence field is equal to its own configuration 1958 sequence number. If they do not match, it MUST generate an error, 1959 either Error_Config_Too_Old or Error_Config_Too_New. In addition, if 1960 the configuration file in the request is too old, it MUST generate a 1961 ConfigUpdate message to update the requesting node. This allows new 1962 configuration documents to propagate quickly throughout the system. 1963 The one exception to this rule is that if the configuration_sequence 1964 field is equal to 0xffff, and the message type is ConfigUpdate, then 1965 the message MUST be accepted regardless of the receiving node's 1966 configuration sequence number. Since 65535 is a special value, peers 1967 sending a new configuration when the configuration sequence is 1968 currently 65534 MUST set the configuration sequence number to 0 when 1969 they send out a new configuration. 1971 5.3.2.2. Destination and Via Lists 1973 The destination list and via lists are sequences of Destination 1974 values: 1976 enum {reserved(0), node(1), resource(2), compressed(3), 1977 /* 128-255 not allowed */ (255) } 1978 DestinationType; 1980 select (destination_type) { 1981 case node: 1982 NodeId node_id; 1984 case resource: 1985 ResourceId resource_id; 1987 case compressed: 1988 opaque compressed_id<0..2^8-1>; 1990 /* This structure may be extended with new types */ 1991 } DestinationData; 1993 struct { 1994 DestinationType type; 1995 uint8 length; 1996 DestinationData destination_data; 1997 } Destination; 1999 struct { 2000 uint16 compressed_id; /* top bit MUST be 1 */ 2001 } Destination; 2003 If a destination structure has its first bit set to 1, then it is a 2004 16 bit integer. If the first bit is not set, then it is a structure 2005 starting with DestinationType. If it is a 16 bit integer, it is 2006 treated as if it were a full structure with a DestinationType of 2007 compressed and a compressed_id that was 2 bytes long with the value 2008 of the 16 bit integer. When the destination structure is not a 16 2009 bit integer, it is the TLV structure with the following contents: 2011 type 2012 The type of the DestinationData Payload Data Unit (PDU). This may 2013 be one of "node", "resource", or "compressed". 2015 length 2016 The length of the destination_data. 2018 destination_data 2019 The destination value itself, which is an encoded DestinationData 2020 structure, depending on the value of "type". 2022 Note: This structure encodes a type, length, value. The length 2023 field specifies the length of the DestinationData values, which 2024 allows the addition of new DestinationTypes. This allows an 2025 implementation which does not understand a given DestinationType 2026 to skip over it. 2028 A DestinationData can be one of three types: 2030 node 2031 A Node-ID. 2033 compressed 2034 A compressed list of Node-IDs and/or resources. Because this 2035 value was compressed by one of the peers, it is only meaningful to 2036 that peer and cannot be decoded by other peers. Thus, it is 2037 represented as an opaque string. 2039 resource 2040 The Resource-ID of the resource which is desired. This type MUST 2041 only appear in the final location of a destination list and MUST 2042 NOT appear in a via list. It is meaningless to try to route 2043 through a resource. 2045 One possible encoding of the 16 bit integer version as an opaque 2046 identifier is to encode an index into a connection table. To avoid 2047 misrouting responses in the event a response is delayed and the 2048 connection table entry has changed, the identifier SHOULD be split 2049 between an index and a generation counter for that index. At 2050 startup, the generation counters should be initialized to random 2051 values. An implementation could use 12 bits for the connection table 2052 index and 3 bits for the generation counter. (Note that this does 2053 not suggest a 4096 entry connection table for every node, only the 2054 ability to encode for a larger connection table.) When a connection 2055 table slot is used for a new connection, the generation counter is 2056 incremented (with wrapping). Connection table slots are used on a 2057 rotating basis to maximize the time interval between uses of the same 2058 slot for different connections. When routing a message to an entry 2059 in the destination list encoding a connection table entry, the node 2060 confirms that the generation counter matches the current generation 2061 counter of that index before forwarding the message. If it does not 2062 match, the message is silently dropped. 2064 5.3.2.3. Forwarding Options 2066 The Forwarding header can be extended with forwarding header options, 2067 which are a series of ForwardingOptions structures: 2069 enum { reservedForwarding(0), (255) } 2070 ForwardingOptionsType; 2072 struct { 2073 ForwardingOptionsType type; 2074 uint8 flags; 2075 uint16 length; 2076 select (type) { 2077 /* This type may be extended */ 2078 } option; 2079 } ForwardingOption; 2081 Each ForwardingOption consists of the following values: 2083 type 2084 The type of the option. This structure allows for unknown options 2085 types. 2087 length 2088 The length of the rest of the structure. 2090 flags 2091 Three flags are defined FORWARD_CRITICAL(0x01), 2092 DESTINATION_CRITICAL(0x02), and RESPONSE_COPY(0x04). These flags 2093 MUST NOT be set in a response. If the FORWARD_CRITICAL flag is 2094 set, any node that would forward the message but does not 2095 understand this options MUST reject the request with an 2096 Error_Unsupported_Forwarding_Option error response. If the 2097 DESTINATION_CRITICAL flag is set, any node that generates a 2098 response to the message but does not understand the forwarding 2099 option MUST reject the request with an 2100 Error_Unsupported_Forwarding_Option error response. If the 2101 RESPONSE_COPY flag is set, any node generating a response MUST 2102 copy the option from the request to the response except that the 2103 RESPONSE_COPY, FORWARD_CRITICAL and DESTINATION_CRITICAL flags 2104 must be cleared. 2106 option 2107 The option value. 2109 5.3.3. Message Contents Format 2111 The second major part of a RELOAD message is the contents part, which 2112 is defined by MessageContents: 2114 enum { reservedMessagesExtension(0), (2^16-1) } MessageExtensionType; 2116 struct { 2117 MessageExtensionType type; 2118 Boolean critical; 2119 opaque extension_contents<0..2^32-1>; 2120 } MessageExtension; 2122 struct { 2123 uint16 message_code; 2124 opaque message_body<0..2^32-1>; 2125 MessageExtensions extensions<0..2^32-1>; 2126 } MessageContents; 2128 The contents of this structure are as follows: 2130 message_code 2131 This indicates the message that is being sent. The code space is 2132 broken up as follows. 2134 0 Reserved 2136 1 .. 0x7fff Requests and responses. These code points are always 2137 paired, with requests being odd and the corresponding response 2138 being the request code plus 1. Thus, "probe_request" (the 2139 Probe request) has value 1 and "probe_answer" (the Probe 2140 response) has value 2 2142 0xffff Error 2143 The message codes are defined in Section 13.8 2144 message_body 2145 The message body itself, represented as a variable-length string 2146 of bytes. The bytes themselves are dependent on the code value. 2147 See the sections describing the various RELOAD methods (Join, 2148 Update, Attach, Store, Fetch, etc.) for the definitions of the 2149 payload contents. 2151 extensions 2152 Extensions to the message. Currently no extensions are defined, 2153 but new extensions can be defined by the process described in 2154 Section 13.14. 2156 All extensions have the following form: 2158 type 2159 The extension type. 2161 critical 2162 Whether this extension must be understood in order to process the 2163 message. If critical = True and the recipient does not understand 2164 the message, it MUST generate an Error_Unknown_Extension error. 2165 If critical = False, the recipient MAY choose to process the 2166 message even if it does not understand the extension. 2168 extension_contents 2169 The contents of the extension (extension-dependent). 2171 5.3.3.1. Response Codes and Response Errors 2173 A peer processing a request returns its status in the message_code 2174 field. If the request was a success, then the message code is the 2175 response code that matches the request (i.e., the next code up). The 2176 response payload is then as defined in the request/response 2177 descriptions. 2179 If the request has failed, then the message code is set to 0xffff 2180 (error) and the payload MUST be an error_response PDU, as shown 2181 below. 2183 When the message code is 0xffff, the payload MUST be an 2184 ErrorResponse. 2186 public struct { 2187 uint16 error_code; 2188 opaque error_info<0..2^16-1>; 2189 } ErrorResponse; 2191 The contents of this structure are as follows: 2193 error_code 2194 A numeric error code indicating the error that occurred. 2196 error_info 2197 An optional arbitrary byte string. Unless otherwise specified, 2198 this will be a UTF-8 text string providing further information 2199 about what went wrong. 2201 The following error code values are defined. The numeric values for 2202 these are defined in Section 13.9. 2204 Error_Forbidden: The requesting node does not have permission to 2205 make this request. 2207 Error_Not_Found: The resource or peer cannot be found or does not 2208 exist. 2210 Error_Request_Timeout: A response to the request has not been 2211 received in a suitable amount of time. The requesting node MAY 2212 resend the request at a later time. 2214 Error_Data_Too_Old: A store cannot be completed because the 2215 storage_time precedes the existing value. 2217 Error_Data_Too_Old: A store cannot be completed because the 2218 storage_time precedes the existing value. 2220 Error_Data_Too_Large: A store cannot be completed because the 2221 requested object exceeds the size limits for that kind. 2223 Error_Generation_Counter_Too_Low: A store cannot be completed 2224 because the generation counter precedes the existing value. 2226 Error_Incompatible_with_Overlay: A peer receiving the request is 2227 using a different overlay, overlay algorithm, or hash algorithm. 2229 Error_Unsupported_Forwarding_Option: A peer receiving the request 2230 with a forwarding options flagged as critical but the peer does 2231 not support this option. See section Section 5.3.2.3. 2233 Error_TTL_Exceeded: A peer receiving the request where the TTL got 2234 decremented to zero. See section Section 5.3.2. 2236 Error_Message_Too_Large: A peer receiving the request that was too 2237 large. See section Section 5.6. 2239 Error_Response_Too_Large: A peer would have generated a response 2240 that is too large per the max_response_length field. 2242 Error_Config_Too_Old: A destination peer received a request with a 2243 configuration sequence that's too old. See Section 5.3.2.1. 2245 Error_Config_Too_New: A destination node received a request with a 2246 configuration sequence that's too new. See Section 5.3.2.1. 2248 Error_Unknown_Kind: A destination node received a request with an 2249 unknown kind-id. See Section 6.4.1.2. 2251 Error_In_Progress: An Attach is already in progress to this peer. 2252 See Section 5.5.1.2. 2254 Error_Unknown_Extension: A destination node received a request with 2255 an unknown extension. 2257 5.3.4. Security Block 2259 The third part of a RELOAD message is the security block. The 2260 security block is represented by a SecurityBlock structure: 2262 struct { 2263 CertificateType type; 2264 opaque certificate<0..2^16-1>; 2265 } GenericCertificate; 2267 struct { 2268 GenericCertificate certificates<0..2^16-1>; 2269 Signature signature; 2270 } SecurityBlock; 2272 The contents of this structure are: 2274 certificates 2275 A bucket of certificates. 2277 signature 2278 A signature over the message contents. 2280 The certificates bucket SHOULD contain all the certificates necessary 2281 to verify every signature in both the message and the internal 2282 message objects. This is the only location in the message which 2283 contains certificates, thus allowing for only a single copy of each 2284 certificate to be sent. In systems which have some alternate 2285 certificate distribution mechanism, some certificates MAY be omitted. 2286 However, implementors should note that this creates the possibility 2287 that messages may not be immediately verifiable because certificates 2288 must first be retrieved. 2290 Each certificate is represented by a GenericCertificate structure, 2291 which has the following contents: 2293 type 2294 The type of the certificate, as defined in [RFC6091]. Only the 2295 use of X.509 certificates is defined in this draft. 2297 certificate 2298 The encoded version of the certificate. For X.509 certificates, 2299 it is the DER form. 2301 The signature is computed over the payload and parts of the 2302 forwarding header. The payload, in case of a Store, may contain an 2303 additional signature computed over a StoreReq structure. All 2304 signatures are formatted using the Signature element. This element 2305 is also used in other contexts where signatures are needed. The 2306 input structure to the signature computation varies depending on the 2307 data element being signed. 2309 enum { reservedSignerIdentity(0), 2310 cert_hash(1), cert_hash_node_id(2), 2311 none(3) 2312 (255)} SignerIdentityType; 2314 struct { 2315 select (identity_type) { 2317 case cert_hash; 2318 HashAlgorithm hash_alg; // From TLS 2319 opaque certificate_hash<0..2^8-1>; 2321 case cert_hash_node_id: 2322 HashAlgorithm hash_alg; // From TLS 2323 opaque certificate_node_id_hash<0..2^8-1>; 2325 case none: 2326 /* empty */ 2327 /* This structure may be extended with new types if necessary*/ 2328 }; 2329 } SignerIdentityValue; 2331 struct { 2332 SignerIdentityType identity_type; 2333 uint16 length; 2334 SignerIdentityValue identity[SignerIdentity.length]; 2335 } SignerIdentity; 2337 struct { 2338 SignatureAndHashAlgorithm algorithm; // From TLS 2339 SignerIdentity identity; 2340 opaque signature_value<0..2^16-1>; 2341 } Signature; 2343 The signature construct contains the following values: 2345 algorithm 2346 The signature algorithm in use. The algorithm definitions are 2347 found in the IANA TLS SignatureAlgorithm Registry and 2348 HashAlgorithm registries. All implementations MUST support 2349 RSASSA-PKCS1-v1_5 [RFC3447] signatures with SHA-256 hashes. 2351 identity 2352 The identity used to form the signature. 2354 signature_value 2355 The value of the signature. 2357 There are two permitted identity formats, one for a certificate with 2358 only one node-id and one for a certificate with multiple node-ids. 2359 In the first case, the cert_hash type SHOULD be used. The hash_alg 2360 field is used to indicate the algorithm used to produce the hash. 2361 The certificate_hash contains the hash of the certificate object 2362 (i.e., the DER-encoded certificate). 2364 In the second case, the cert_hash_node_id type MUST be used. The 2365 hash_alg is as in cert_hash but the cert_hash_node_id is computed 2366 over the NodeId used to sign concatenated with the certificate. 2367 I.e., H(NodeID || certificate). The NodeId is represented without 2368 any framing or length fields, as simple raw bytes. This is safe 2369 because NodeIds are fixed-length for a given overlay. 2371 For signatures over messages the input to the signature is computed 2372 over: 2374 overlay || transaction_id || MessageContents || SignerIdentity 2376 where overlay and transaction_id come from the forwarding header and 2377 || indicates concatenation. 2379 The input to signatures over data values is different, and is 2380 described in Section 6.1. 2382 All RELOAD messages MUST be signed. Upon receipt, the receiving node 2383 MUST verify the signature and the authorizing certificate. This 2384 check provides a minimal level of assurance that the sending node is 2385 a valid part of the overlay as well as cryptographic authentication 2386 of the sending node. In addition, responses MUST be checked as 2387 follows: 2389 1. The response to a message sent to a specific Node-ID MUST have 2390 been sent by that Node-ID. 2391 2. The response to a message sent to a Resource-Id MUST have been 2392 sent by a Node-ID which is as close to or closer to the target 2393 Resource-Id than any node in the requesting node's neighbor 2394 table. 2396 The second condition serves as a primitive check for responses from 2397 wildly wrong nodes but is not a complete check. Note that in periods 2398 of churn, it is possible for the requesting node to obtain a closer 2399 neighbor while the request is outstanding. This will cause the 2400 response to be rejected and the request to be retransmitted. 2402 In addition, some methods (especially Store) have additional 2403 authentication requirements, which are described in the sections 2404 covering those methods. 2406 5.4. Overlay Topology 2408 As discussed in previous sections, RELOAD does not itself implement 2409 any overlay topology. Rather, it relies on Topology Plugins, which 2410 allow a variety of overlay algorithms to be used while maintaining 2411 the same RELOAD core. This section describes the requirements for 2412 new topology plugins and the methods that RELOAD provides for overlay 2413 topology maintenance. 2415 5.4.1. Topology Plugin Requirements 2417 When specifying a new overlay algorithm, at least the following need 2418 to be described: 2420 o Joining procedures, including the contents of the Join message. 2421 o Stabilization procedures, including the contents of the Update 2422 message, the frequency of topology probes and keepalives, and the 2423 mechanism used to detect when peers have disconnected. 2424 o Exit procedures, including the contents of the Leave message. 2425 o The length of the Resource-IDs. For DHTs, the hash algorithm to 2426 compute the hash of an identifier. 2427 o The procedures that peers use to route messages. 2428 o The replication strategy used to ensure data redundancy. 2430 All overlay algorithms MUST specify maintenance procedures that send 2431 Updates to clients and peers that have established connections to the 2432 peer responsible for a particular ID when the responsibility for that 2433 ID changes. Because tracking this information is difficult, overlay 2434 algorithms MAY simply specify that an Update is sent to all members 2435 of the Connection Table whenever the range of IDs for which the peer 2436 is responsible changes. 2438 5.4.2. Methods and types for use by topology plugins 2440 This section describes the methods that topology plugins use to join, 2441 leave, and maintain the overlay. 2443 5.4.2.1. Join 2445 A new peer (but one that already has credentials) uses the JoinReq 2446 message to join the overlay. The JoinReq is sent to the responsible 2447 peer depending on the routing mechanism described in the topology 2448 plugin. This notifies the responsible peer that the new peer is 2449 taking over some of the overlay and it needs to synchronize its 2450 state. 2452 struct { 2453 NodeId joining_peer_id; 2454 opaque overlay_specific_data<0..2^16-1>; 2455 } JoinReq; 2457 The minimal JoinReq contains only the Node-ID which the sending peer 2458 wishes to assume. Overlay algorithms MAY specify other data to 2459 appear in this request. Receivers of the JoinReq MUST verify that 2460 the joining_peer_id field matches the Node-ID used to sign the 2461 message and if not MUST reject the message with an Error_Forbidden 2462 error. 2464 If the request succeeds, the responding peer responds with a JoinAns 2465 message, as defined below: 2467 struct { 2468 opaque overlay_specific_data<0..2^16-1>; 2469 } JoinAns; 2471 If the request succeeds, the responding peer MUST follow up by 2472 executing the right sequence of Stores and Updates to transfer the 2473 appropriate section of the overlay space to the joining peer. In 2474 addition, overlay algorithms MAY define data to appear in the 2475 response payload that provides additional info. 2477 In general, nodes which cannot form connections SHOULD report an 2478 error. However, implementations MUST provide some mechanism whereby 2479 nodes can determine that they are potentially the first node and take 2480 responsibility for the overlay. This specification does not mandate 2481 any particular mechanism, but a configuration flag or setting seems 2482 appropriate. 2484 5.4.2.2. Leave 2486 The LeaveReq message is used to indicate that a node is exiting the 2487 overlay. A node SHOULD send this message to each peer with which it 2488 is directly connected prior to exiting the overlay. 2490 struct { 2491 NodeId leaving_peer_id; 2492 opaque overlay_specific_data<0..2^16-1>; 2493 } LeaveReq; 2495 LeaveReq contains only the Node-ID of the leaving peer. Overlay 2496 algorithms MAY specify other data to appear in this request. 2497 Receivers of the LeaveReq MUST verify that the leaving_peer_id field 2498 matches the Node-ID used to sign the message and if not MUST reject 2499 the message with an Error_Forbidden error. 2501 Upon receiving a Leave request, a peer MUST update its own routing 2502 table, and send the appropriate Store/Update sequences to re- 2503 stabilize the overlay. 2505 5.4.2.3. Update 2507 Update is the primary overlay-specific maintenance message. It is 2508 used by the sender to notify the recipient of the sender's view of 2509 the current state of the overlay (its routing state), and it is up to 2510 the recipient to take whatever actions are appropriate to deal with 2511 the state change. In general, peers send Update messages to all 2512 their adjacencies whenever they detect a topology shift. 2514 When a peer receives an Attach request with the send_update flag set 2515 to "true" (Section 5.4.2.4.1, it MUST send an Update message back to 2516 the sender of the Attach request after the completion of the 2517 corresponding ICE check and TLS connection. Note that the sender of 2518 a such Attach request may not have joined the overlay yet. 2520 When a peer detects through an Update that it is no longer 2521 responsible for any data value it is storing, it MUST attempt to 2522 Store a copy to the correct node unless it knows the newly 2523 responsible node already has a copy of the data. This prevents data 2524 loss during large-scale topology shifts such as the merging of 2525 partitioned overlays. 2527 The contents of the UpdateReq message are completely overlay- 2528 specific. The UpdateAns response is expected to be either success or 2529 an error. 2531 5.4.2.4. RouteQuery 2533 The RouteQuery request allows the sender to ask a peer where they 2534 would route a message directed to a given destination. In other 2535 words, a RouteQuery for a destination X requests the Node-ID for the 2536 node that the receiving peer would next route to in order to get to 2537 X. A RouteQuery can also request that the receiving peer initiate an 2538 Update request to transfer the receiving peer's routing table. 2540 One important use of the RouteQuery request is to support iterative 2541 routing. The sender selects one of the peers in its routing table 2542 and sends it a RouteQuery message with the destination_object set to 2543 the Node-ID or Resource-ID it wishes to route to. The receiving peer 2544 responds with information about the peers to which the request would 2545 be routed. The sending peer MAY then use the Attach method to attach 2546 to that peer(s), and repeat the RouteQuery. Eventually, the sender 2547 gets a response from a peer that is closest to the identifier in the 2548 destination_object as determined by the topology plugin. At that 2549 point, the sender can send messages directly to that peer. 2551 5.4.2.4.1. Request Definition 2553 A RouteQueryReq message indicates the peer or resource that the 2554 requesting node is interested in. It also contains a "send_update" 2555 option allowing the requesting node to request a full copy of the 2556 other peer's routing table. 2558 struct { 2559 Boolean send_update; 2560 Destination destination; 2561 opaque overlay_specific_data<0..2^16-1>; 2562 } RouteQueryReq; 2564 The contents of the RouteQueryReq message are as follows: 2566 send_update 2567 A single byte. This may be set to "true" to indicate that the 2568 requester wishes the responder to initiate an Update request 2569 immediately. Otherwise, this value MUST be set to "false". 2571 destination 2572 The destination which the requester is interested in. This may be 2573 any valid destination object, including a Node-ID, compressed ids, 2574 or Resource-ID. 2576 overlay_specific_data 2577 Other data as appropriate for the overlay. 2579 5.4.2.4.2. Response Definition 2581 A response to a successful RouteQueryReq request is a RouteQueryAns 2582 message. This is completely overlay specific. 2584 5.4.2.5. Probe 2586 Probe provides primitive "exploration" services: it allows node to 2587 determine which resources another node is responsible for; and it 2588 allows some discovery services using multicast, anycast, or 2589 broadcast. A probe can be addressed to a specific Node-ID, or the 2590 peer controlling a given location (by using a Resource-ID). In 2591 either case, the target Node-IDs respond with a simple response 2592 containing some status information. 2594 5.4.2.5.1. Request Definition 2596 The ProbeReq message contains a list (potentially empty) of the 2597 pieces of status information that the requester would like the 2598 responder to provide. 2600 enum { reservedProbeInformation(0), responsible_set(1), 2601 num_resources(2), uptime(3), (255)} 2602 ProbeInformationType; 2604 struct { 2605 ProbeInformationType requested_info<0..2^8-1>; 2606 } ProbeReq 2608 The currently defined values for ProbeInformation are: 2610 responsible_set 2611 indicates that the peer should Respond with the fraction of the 2612 overlay for which the responding peer is responsible. 2614 num_resources 2615 indicates that the peer should Respond with the number of 2616 resources currently being stored by the peer. 2618 uptime 2619 indicates that the peer should Respond with how long the peer has 2620 been up in seconds. 2622 5.4.2.5.2. Response Definition 2624 A successful ProbeAns response contains the information elements 2625 requested by the peer. 2627 struct { 2628 select (type) { 2629 case responsible_set: 2630 uint32 responsible_ppb; 2632 case num_resources: 2633 uint32 num_resources; 2635 case uptime: 2636 uint32 uptime; 2637 /* This type may be extended */ 2639 }; 2640 } ProbeInformationData; 2642 struct { 2643 ProbeInformationType type; 2644 uint8 length; 2645 ProbeInformationData value; 2646 } ProbeInformation; 2648 struct { 2649 ProbeInformation probe_info<0..2^16-1>; 2650 } ProbeAns; 2652 A ProbeAns message contains a sequence of ProbeInformation 2653 structures. Each has a "length" indicating the length of the 2654 following value field. This structure allows for unknown option 2655 types. 2657 Each of the current possible Probe information types is a 32-bit 2658 unsigned integer. For type "responsible_ppb", it is the fraction of 2659 the overlay for which the peer is responsible in parts per billion. 2660 For type "num_resources", it is the number of resources the peer is 2661 storing. For the type "uptime" it is the number of seconds the peer 2662 has been up. 2664 The responding peer SHOULD include any values that the requesting 2665 node requested and that it recognizes. They SHOULD be returned in 2666 the requested order. Any other values MUST NOT be returned. 2668 5.5. Forwarding and Link Management Layer 2670 Each node maintains connections to a set of other nodes defined by 2671 the topology plugin. This section defines the methods RELOAD uses to 2672 form and maintain connections between nodes in the overlay. Three 2673 methods are defined: 2675 Attach: used to form RELOAD connections between nodes using ICE 2676 for NAT traversal. When node A wants to connect to node B, it 2677 sends an Attach message to node B through the overlay. The Attach 2678 contains A's ICE parameters. B responds with its ICE parameters 2679 and the two nodes perform ICE to form connection. Attach also 2680 allows two nodes to connect via No-ICE instead of full ICE. 2681 AppAttach: used to form application layer connections between 2682 nodes. 2683 Ping: is a simple request/response which is used to verify 2684 connectivity of the target peer. 2686 5.5.1. Attach 2688 A node sends an Attach request when it wishes to establish a direct 2689 TCP or UDP connection to another node for the purpose of sending 2690 RELOAD messages. A client that can establish a connection directly 2691 need not send an attach as described in the second bullet of 2692 Section 3.2.1 2694 As described in Section 5.1, an Attach may be routed to either a 2695 Node-ID or to a Resource-ID. An Attach routed to a specific Node-ID 2696 will fail if that node is not reached. An Attach routed to a 2697 Resource-ID will establish a connection with the peer currently 2698 responsible for that Resource-ID, which may be useful in establishing 2699 a direct connection to the responsible peer for use with frequent or 2700 large resource updates. 2702 An Attach in and of itself does not result in updating the routing 2703 table of either node. That function is performed by Updates. If 2704 node A has Attached to node B, but not received any Updates from B, 2705 it MAY route messages which are directly addressed to B through that 2706 channel but MUST NOT route messages through B to other peers via that 2707 channel. The process of Attaching is separate from the process of 2708 becoming a peer (using Join and Update), to prevent half-open states 2709 where a node has started to form connections but is not really ready 2710 to act as a peer. Thus, clients (unlike peers) can simply Attach 2711 without sending Join or Update. 2713 5.5.1.1. Request Definition 2715 An Attach request message contains the requesting node ICE connection 2716 parameters formatted into a binary structure. 2718 enum { reservedOverlayLink(0), DTLS-UDP-SR(1), 2719 DTLS-UDP-SR-NO-ICE(3), TLS-TCP-FH-NO-ICE(4), 2720 (255) } OverlayLinkType; 2722 enum { reservedCand(0), host(1), srflx(2), prflx(3), relay(4), 2723 (255) } CandType; 2725 struct { 2726 opaque name<0..2^16-1>; 2727 opaque value<0..2^16-1>; 2728 } IceExtension; 2730 struct { 2731 IpAddressPort addr_port; 2732 OverlayLinkType overlay_link; 2733 opaque foundation<0..255>; 2734 uint32 priority; 2735 CandType type; 2736 select (type){ 2737 case host: 2738 ; /* Nothing */ 2739 case srflx: 2740 case prflx: 2741 case relay: 2742 IpAddressPort rel_addr_port; 2743 }; 2744 IceExtension extensions<0..2^16-1>; 2745 } IceCandidate; 2747 struct { 2748 opaque ufrag<0..2^8-1>; 2749 opaque password<0..2^8-1>; 2750 opaque role<0..2^8-1>; 2751 IceCandidate candidates<0..2^16-1>; 2752 Boolean send_update; 2753 } AttachReqAns; 2755 The values contained in AttachReqAns are: 2757 ufrag 2758 The username fragment (from ICE). 2760 password 2761 The ICE password. 2763 role 2764 An active/passive/actpass attribute from RFC 4145 [RFC4145]. This 2765 value MUST be 'passive' for the offerer (the peer sending the 2766 Attach request) and 'active' for the answerer (the peer sending 2767 the Attach response). 2769 candidates 2770 One or more ICE candidate values, as described below. 2771 send_update 2772 Has the same meaning as the send_update field in RouteQueryReq. 2774 Each ICE candidate is represented as an IceCandidate structure, which 2775 is a direct translation of the information from the ICE string 2776 structures, with the exception of the component ID. Since there is 2777 only one component, it is always 1, and thus left out of the PDU. 2778 The remaining values are specified as follows: 2780 addr_port 2781 corresponds to the connection-address and port productions. 2783 overlay_link 2784 corresponds to the OverlayLinkType production, Overlay Link 2785 protocols used with No-ICE MUST specify "No-ICE" in their 2786 description. Future overlay link values can be added be defining 2787 new OverlayLinkType values in the IANA registry in Section 13.10. 2788 Future extensions to the encapsulation or framing that provide for 2789 backward compatibility with that specified by a previously defined 2790 OverlayLinkType values MUST use that previous value. 2791 OverlayLinkType protocols are defined in Section 5.6 2792 A single AttachReqAns MUST NOT include both candidates whose 2793 OverlayLinkType protocols use ICE (the default) and candidates 2794 that specify "No-ICE". 2796 foundation 2797 corresponds to the foundation production. 2799 priority 2800 corresponds to the priority production. 2802 type 2803 corresponds to the cand-type production. 2805 rel_addr_port 2806 corresponds to the rel-addr and rel-port productions. Only 2807 present for type "relay". 2809 extensions 2810 ICE extensions. The name and value fields correspond to binary 2811 translations of the equivalent fields in the ICE extensions. 2813 These values should be generated using the procedures described in 2814 Section 5.5.1.3. 2816 5.5.1.2. Response Definition 2818 If a peer receives an Attach request, it MUST determine how to 2819 process the request as follows: 2821 o If it has not initiated an Attach request to the originating peer 2822 of this Attach request, it MUST process this request and SHOULD 2823 generate its own response with an AttachReqAns. It should then 2824 begin ICE checks. 2825 o If it has already sent an Attach request to and received the 2826 response from the originating peer of this Attach request, and as 2827 a as a result, an ICE check and TLS connection is in progress, 2828 then it SHOULD generate an Error_In_Progress error instead of an 2829 AttachReqAns. 2830 o If it has already sent an Attach request to but not yet received 2831 the response from the originating peer of this Attach request, it 2832 SHOULD apply the following tie-breaker heuristic to determine how 2833 to handle this Attach request and the incomplete Attach request it 2834 has sent out: 2835 * If the peer's own Node-ID is smaller when compared as big- 2836 endian unsigned integers, it MUST cancel its own incomplete 2837 Attach request. It MUST then process this Attach request, 2838 generate an AttachReqAns response, and proceed with the 2839 corresponding ICE check. 2840 * If the peer's own Node-ID is larger when compared as big-endien 2841 unsigned integers, it MUST generate an Error_In_Progress error 2842 to this Attach request, then proceed to wait for and complete 2843 the Attach and the corresponding ICE check it has originated. 2844 o If the peer is overloaded or detects some other kind of error, it 2845 MAY generate an error instead of an AttachReqAns. 2847 When a peer receives an Attach response, it SHOULD parse the response 2848 and begin its own ICE checks. 2850 5.5.1.3. Using ICE With RELOAD 2852 This section describes the profile of ICE that is used with RELOAD. 2853 RELOAD implementations MUST implement full ICE. 2855 In ICE as defined by [RFC5245], SDP is used to carry the ICE 2856 parameters. In RELOAD, this function is performed by a binary 2857 encoding in the Attach method. This encoding is more restricted than 2858 the SDP encoding because the RELOAD environment is simpler: 2860 o Only a single media stream is supported. 2861 o In this case, the "stream" refers not to RTP or other types of 2862 media, but rather to a connection for RELOAD itself or other 2863 application-layer protocols such as SIP. 2864 o RELOAD only allows for a single offer/answer exchange. Unlike the 2865 usage of ICE within SIP, there is never a need to send a 2866 subsequent offer to update the default candidates to match the 2867 ones selected by ICE. 2869 An agent follows the ICE specification as described in [RFC5245] with 2870 the changes and additional procedures described in the subsections 2871 below. 2873 5.5.1.4. Collecting STUN Servers 2875 ICE relies on the node having one or more STUN servers to use. In 2876 conventional ICE, it is assumed that nodes are configured with one or 2877 more STUN servers through some out of band mechanism. This is still 2878 possible in RELOAD but RELOAD also learns STUN servers as it connects 2879 to other peers. Because all RELOAD peers implement ICE and use STUN 2880 keepalives, every peer is a capable of responding to STUN Binding 2881 requests [RFC5389]. Accordingly, any peer that a node knows about 2882 can be used like a STUN server -- though of course it may be behind a 2883 NAT. 2885 A peer on a well-provisioned wide-area overlay will be configured 2886 with one or more bootstrap nodes. These nodes make an initial list 2887 of STUN servers. However, as the peer forms connections with 2888 additional peers, it builds more peers it can use like STUN servers. 2890 Because complicated NAT topologies are possible, a peer may need more 2891 than one STUN server. Specifically, a peer that is behind a single 2892 NAT will typically observe only two IP addresses in its STUN checks: 2893 its local address and its server reflexive address from a STUN server 2894 outside its NAT. However, if there are more NATs involved, it may 2895 learn additional server reflexive addresses (which vary based on 2896 where in the topology the STUN server is). To maximize the chance of 2897 achieving a direct connection, a peer SHOULD group other peers by the 2898 peer-reflexive addresses it discovers through them. It SHOULD then 2899 select one peer from each group to use as a STUN server for future 2900 connections. 2902 Only peers to which the peer currently has connections may be used. 2903 If the connection to that host is lost, it MUST be removed from the 2904 list of stun servers and a new server from the same group MUST be 2905 selected unless there are no others servers in the group in which 2906 case some other peer MAY be used. 2908 5.5.1.5. Gathering Candidates 2910 When a node wishes to establish a connection for the purposes of 2911 RELOAD signaling or application signaling, it follows the process of 2912 gathering candidates as described in Section 4 of ICE [RFC5245]. 2913 RELOAD utilizes a single component. Consequently, gathering for 2914 these "streams" requires a single component. In the case where a 2915 node has not yet found a TURN server, the agent would not include a 2916 relayed candidate. 2918 The ICE specification assumes that an ICE agent is configured with, 2919 or somehow knows of, TURN and STUN servers. RELOAD provides a way 2920 for an agent to learn these by querying the overlay, as described in 2921 Section 5.5.1.4 and Section 8. 2923 The default candidate selection described in Section 4.1.4 of ICE is 2924 ignored; defaults are not signaled or utilized by RELOAD. 2926 An alternative to using the full ICE supported by the Attach request 2927 is to use No-ICE mechanism by providing candidates with "No-ICE" 2928 Overlay Link protocols. Configuration for the overlay indicates 2929 whether or not these Overlay Link protocols can be used. An overlay 2930 MUST be either all ICE or all No-ICE. 2932 No-ICE will not work in all of the scenarios where ICE would work, 2933 but in some cases, particularly those with no NATs or firewalls, it 2934 will work. 2936 5.5.1.6. Prioritizing Candidates 2938 However, standardization of additional protocols for use with ICE is 2939 expected, including TCP[I-D.ietf-mmusic-ice-tcp] and protocols such 2940 as SCTP and DCCP. UDP encapsulations for SCTP and DCCP would expand 2941 the available Overlay Link protocols available for RELOAD. When 2942 additional protocols are available, the following prioritization is 2943 RECOMMENDED: 2945 o Highest priority is assigned to protocols that offer well- 2946 understood congestion and flow control without head of line 2947 blocking. For example, SCTP without message ordering, DCCP, or 2948 those protocols encapsulated using UDP. 2949 o Second highest priority is assigned to protocols that offer well- 2950 understood congestion and flow control but have head of line 2951 blocking such as TCP. 2952 o Lowest priority is assigned to protocols encapsulated over UDP 2953 that do not implement well-established congestion control 2954 algorithms. The DTLS/UDP with SR overlay link protocol is an 2955 example of such a protocol. 2957 5.5.1.7. Encoding the Attach Message 2959 Section 4.3 of ICE describes procedures for encoding the SDP for 2960 conveying RELOAD candidates. Instead of actually encoding an SDP, 2961 the candidate information (IP address and port and transport 2962 protocol, priority, foundation, type and related address) is carried 2963 within the attributes of the Attach request or its response. 2964 Similarly, the username fragment and password are carried in the 2965 Attach message or its response. Section 5.5.1 describes the detailed 2966 attribute encoding for Attach. The Attach request and its response 2967 do not contain any default candidates or the ice-lite attribute, as 2968 these features of ICE are not used by RELOAD. 2970 Since the Attach request contains the candidate information and short 2971 term credentials, it is considered as an offer for a single media 2972 stream that happens to be encoded in a format different than SDP, but 2973 is otherwise considered a valid offer for the purposes of following 2974 the ICE specification. Similarly, the Attach response is considered 2975 a valid answer for the purposes of following the ICE specification. 2977 5.5.1.8. Verifying ICE Support 2979 An agent MUST skip the verification procedures in Section 5.1 and 6.1 2980 of ICE. Since RELOAD requires full ICE from all agents, this check 2981 is not required. 2983 5.5.1.9. Role Determination 2985 The roles of controlling and controlled as described in Section 5.2 2986 of ICE are still utilized with RELOAD. However, the offerer (the 2987 entity sending the Attach request) will always be controlling, and 2988 the answerer (the entity sending the Attach response) will always be 2989 controlled. The connectivity checks MUST still contain the ICE- 2990 CONTROLLED and ICE-CONTROLLING attributes, however, even though the 2991 role reversal capability for which they are defined will never be 2992 needed with RELOAD. This is to allow for a common codebase between 2993 ICE for RELOAD and ICE for SDP. 2995 5.5.1.10. Full ICE 2997 When the overlay uses ICE , connectivity checks and nominations are 2998 used as in regular ICE. 3000 5.5.1.10.1. Connectivity Checks 3002 The processes of forming check lists in Section 5.7 of ICE, 3003 scheduling checks in Section 5.8, and checking connectivity checks in 3004 Section 7 are used with RELOAD without change. 3006 5.5.1.10.2. Concluding ICE 3008 The procedures in Section 8 of ICE are followed to conclude ICE, with 3009 the following exceptions: 3011 o The controlling agent MUST NOT attempt to send an updated offer 3012 once the state of its single media stream reaches Completed. 3013 o Once the state of ICE reaches Completed, the agent can immediately 3014 free all unused candidates. This is because RELOAD does not have 3015 the concept of forking, and thus the three second delay in Section 3016 8.3 of ICE does not apply. 3018 5.5.1.10.3. Media Keepalives 3020 STUN MUST be utilized for the keepalives described in Section 10 of 3021 ICE. 3023 5.5.1.11. No-ICE 3025 No-ICE is selected when either side has provided "no ICE" Overlay 3026 Link candidates. STUN is not used for connectivity checks when doing 3027 No-ICE; instead the DTLS or TLS handshake (or similar security layer 3028 of future overlay link protocols) forms the connectivity check. The 3029 certificate exchanged during the (D)TLS handshake must match the node 3030 that sent the AttachReqAns and if it does not, the connection MUST be 3031 closed. 3033 5.5.1.12. Subsequent Offers and Answers 3035 An agent MUST NOT send a subsequent offer or answer. Thus, the 3036 procedures in Section 9 of ICE MUST be ignored. 3038 5.5.1.13. Sending Media 3040 The procedures of Section 11 of ICE apply to RELOAD as well. 3041 However, in this case, the "media" takes the form of application 3042 layer protocols (e.g. RELOAD) over TLS or DTLS. Consequently, once 3043 ICE processing completes, the agent will begin TLS or DTLS procedures 3044 to establish a secure connection. The node which sent the Attach 3045 request MUST be the TLS server. The other node MUST be the TLS 3046 client. The server MUST request TLS client authentication. The 3047 nodes MUST verify that the certificate presented in the handshake 3048 matches the identity of the other peer as found in the Attach 3049 message. Once the TLS or DTLS signaling is complete, the application 3050 protocol is free to use the connection. 3052 The concept of a previous selected pair for a component does not 3053 apply to RELOAD, since ICE restarts are not possible with RELOAD. 3055 5.5.1.14. Receiving Media 3057 An agent MUST be prepared to receive packets for the application 3058 protocol (TLS or DTLS carrying RELOAD, SIP or anything else) at any 3059 time. The jitter and RTP considerations in Section 11 of ICE do not 3060 apply to RELOAD. 3062 5.5.2. AppAttach 3064 A node sends an AppAttach request when it wishes to establish a 3065 direct connection to another node for the purposes of sending 3066 application layer messages. AppAttach is nearly identical to Attach, 3067 except for the purpose of the connection: it is used to transport 3068 non-RELOAD "media". A separate request is used to avoid implementor 3069 confusion between the two methods (this was found to be a real 3070 problem with initial implementations). The AppAttach request and its 3071 response contain an application attribute, which indicates what 3072 protocol is to be run over the connection. 3074 5.5.2.1. Request Definition 3076 An AppAttachReq message contains the requesting node's ICE connection 3077 parameters formatted into a binary structure. 3079 struct { 3080 opaque ufrag<0..2^8-1>; 3081 opaque password<0..2^8-1>; 3082 uint16 application; 3083 opaque role<0..2^8-1>; 3084 IceCandidate candidates<0..2^16-1>; 3085 } AppAttachReq; 3087 The values contained in AppAttachReq and AppAttachAns are: 3089 ufrag 3090 The username fragment (from ICE) 3092 password 3093 The ICE password. 3095 application 3096 A 16-bit application-id as defined in the Section 13.5. This 3097 number represents the IANA registered application that is going to 3098 send data on this connection. For SIP, this is 5060 or 5061. 3100 role 3101 An active/passive/actpass attribute from RFC 4145 [RFC4145]. 3103 candidates 3104 One or more ICE candidate values 3106 The application using connection set up with this request is 3107 responsible for providing sufficiently frequent keep traffic for NAT 3108 and Firewall keep alive and for deciding when to close the 3109 connection. 3111 5.5.2.2. Response Definition 3113 If a peer receives an AppAttach request, it SHOULD process the 3114 request and generate its own response with a AppAttachAns. It should 3115 then begin ICE checks. When a peer receives an AppAttach response, 3116 it SHOULD parse the response and begin its own ICE checks. If the 3117 application ID is not supported, the peer MUST reply with an 3118 Error_Not_Found error. 3120 struct { 3121 opaque ufrag<0..2^8-1>; 3122 opaque password<0..2^8-1>; 3123 uint16 application; 3124 opaque role<0..2^8-1>; 3125 IceCandidate candidates<0..2^16-1>; 3126 } AppAttachAns; 3128 The meaning of the fields is the same as in the AppAttachReq. 3130 5.5.3. Ping 3132 Ping is used to test connectivity along a path. A ping can be 3133 addressed to a specific Node-ID, to the peer controlling a given 3134 location (by using a resource ID), or to the broadcast Node-ID 3135 (2^128-1). 3137 5.5.3.1. Request Definition 3139 struct { 3140 opaque<0..2^16-1> padding; 3141 } PingReq 3143 The Ping request is empty of meaningful contents. However, it may 3144 contain up to 65535 bytes of padding to facilitate the discovery of 3145 overlay maximum packet sizes. 3147 5.5.3.2. Response Definition 3149 A successful PingAns response contains the information elements 3150 requested by the peer. 3152 struct { 3153 uint64 response_id; 3154 uint64 time; 3155 } PingAns; 3157 A PingAns message contains the following elements: 3159 response_id 3160 A randomly generated 64-bit response ID. This is used to 3161 distinguish Ping responses. 3163 time 3164 The time when the Ping response was created represented in the 3165 same way as storage_time defined in Section 6. 3167 5.5.4. ConfigUpdate 3169 The ConfigUpdate method is used to push updated configuration data 3170 across the overlay. Whenever a node detects that another node has 3171 old configuration data, it MUST generate a ConfigUpdate request. The 3172 ConfigUpdate request allows updating of two kinds of data: the 3173 configuration data (Section 5.3.2.1) and kind information 3174 (Section 6.4.1.1). 3176 5.5.4.1. Request Definition 3178 enum { reservedConfigUpdate(0), config(1), kind(2), (255) } 3179 ConfigUpdateType; 3181 typedef uint32 KindId; 3182 typedef opaque KindDescription<0..2^16-1>; 3184 struct { 3185 ConfigUpdateType type; 3186 uint32 length; 3188 select (type) { 3189 case config: 3190 opaque config_data<0..2^24-1>; 3192 case kind: 3193 KindDescription kinds<0..2^24-1>; 3195 /* This structure may be extended with new types*/ 3196 }; 3197 } ConfigUpdateReq; 3199 The ConfigUpdateReq message contains the following elements: 3201 type 3202 The type of the contents of the message. This structure allows 3203 for unknown content types. 3204 length 3205 The length of the remainder of the message. This is included to 3206 preserve backward compatibility and is 32 bits instead of 24 to 3207 facilitate easy conversion between network and host byte order. 3208 config_data (type==config) 3209 The contents of the configuration document. 3210 kinds (type==kind) 3211 One or more XML kind-block productions (see Section 10.1). These 3212 MUST be encoded with UTF-8 and assume a default namespace of 3213 "urn:ietf:params:xml:ns:p2p:config-base". 3215 5.5.4.2. Response Definition 3217 struct { 3218 } ConfigUpdateAns 3220 If the ConfigUpdateReq is of type "config" it MUST only be processed 3221 if all the following are true: 3223 o The sequence number in the document is greater than the current 3224 configuration sequence number. 3225 o The configuration document is correctly digitally signed (see 3226 Section 10 for details on signatures. 3227 Otherwise appropriate errors MUST be generated. 3229 If the ConfigUpdateReq is of type "kind" it MUST only be processed if 3230 it is correctly digitally signed by an acceptable kind signer as 3231 specified in the configuration file. Details on kind-signer field in 3232 the configuration file is described in Section 10.1. In addition, if 3233 the kind update conflicts with an existing known kind (i.e., it is 3234 signed by a different signer), then it should be rejected with 3235 "Error_Forbidden". This should not happen in correctly functioning 3236 overlays. 3238 If the update is acceptable, then the node MUST reconfigure itself to 3239 match the new information. This may include adding permissions for 3240 new kinds, deleting old kinds, or even, in extreme circumstances, 3241 exiting and reentering the overlay, if, for instance, the DHT 3242 algorithm has changed. 3244 The response for ConfigUpdate is empty. 3246 5.6. Overlay Link Layer 3248 RELOAD can use multiple Overlay Link protocols to send its messages. 3249 Because ICE is used to establish connections (see Section 5.5.1.3), 3250 RELOAD nodes are able to detect which Overlay Link protocols are 3251 offered by other nodes and establish connections between them. Any 3252 link protocol needs to be able to establish a secure, authenticated 3253 connection and to provide data origin authentication and message 3254 integrity for individual data elements. RELOAD currently supports 3255 three Overlay Link protocols: 3257 o DTLS [RFC4347] over UDP with Simple Reliability (SR) 3258 o TLS [RFC5246] over TCP with Framing Header, No-ICE 3259 o DTLS [RFC4347] over UDP with SR, No-ICE 3261 Note that although UDP does not properly have "connections", both TLS 3262 and DTLS have a handshake which establishes a similar, stateful 3263 association, and we simply refer to these as "connections" for the 3264 purposes of this document. 3266 If a peer receives a message that is larger than value of max- 3267 message-size defined in the overlay configuration, the peer SHOULD 3268 send an Error_Message_Too_Large error and then close the TLS or DTLS 3269 session from which the message was received. Note that this error 3270 can be sent and the session closed before receiving the complete 3271 message. If the forwarding header is larger than the max-message- 3272 size, the receiver SHOULD close the TLS or DTLS session without 3273 sending an error. 3275 The Framing Header (FH) is used to frame messages and provide timing 3276 when used on a reliable stream-based transport protocol. Simple 3277 Reliability (SR) makes use of the FH to provide congestion control 3278 and semi-reliability when using unreliable message-oriented transport 3279 protocols. We will first define each of these algorithms, then 3280 define overlay link protocols that use them. 3282 Note: We expect future Overlay Link protocols to define replacements 3283 for all components of these protocols, including the framing header. 3284 These protocols have been chosen for simplicity of implementation and 3285 reasonable performance. 3287 Note to implementers: There are inherent tradeoffs in utilizing 3288 short timeouts to determine when a link has failed. To balance the 3289 tradeoffs, an implementation should be able to quickly act to remove 3290 entries from the routing table when there is reason to suspect the 3291 link has failed. For example, in a Chord derived overlay algorithm, 3292 a closer finger table entry could be substituted for an entry in the 3293 finger table that has experienced a timeout. That entry can be 3294 restored if it proves to resume functioning, or replaced at some 3295 point in the future if necessary. End-to-end retransmissions will 3296 handle any lost messages, but only if the failing entries do not 3297 remain in the finger table for subsequent retransmissions. 3299 5.6.1. Future Overlay Link Protocols 3301 It is possible to define new link-layer protocols and apply them to a 3302 new overlay using the "overlay-link-protocol" configuration directive 3303 (see Section 10.1.). However, any new protocols MUST meet the 3304 following requirements. 3306 Endpoint authentication When a node forms an association with 3307 another endpoint, it MUST be possible to cryptographically verify 3308 that the endpoint has a given Node-Id. 3310 Traffic origin authentication and integrity When a node receives 3311 traffic from another endpoint, it MUST be possible to 3312 cryptographically verify that the traffic came from a given 3313 association and that it has not been modified in transit from the 3314 other endpoint in the association. The overlay link protocol MUST 3315 also provide replay prevention/detection. 3317 Traffic confidentiality When a node sends traffic to another 3318 endpoint, it MUST NOT be possible for a third party not involved 3319 in the association to determine the contents of that traffic. 3321 Any new overlay protocol MUST be defined via RFC 5226 Standards 3322 Action; see Section 13.11. 3324 5.6.1.1. HIP 3326 In a Host Identity Protocol Based Overlay Networking Environment (HIP 3327 BONE) [RFC6079] HIP [RFC5201] provides connection management (e.g., 3328 NAT traversal and mobility) and security for the overlay network. 3329 The P2PSIP Working Group has expressed interest in supporting a HIP- 3330 based link protocol. Such support would require specifying such 3331 details as: 3333 o How to issue certificates which provided identities meaningful to 3334 the HIP base exchange. We anticipate that this would require a 3335 mapping between ORCHIDs and NodeIds. 3336 o How to carry the HIP I1 and I2 messages. 3337 o How to carry RELOAD messages over HIP. 3339 [I-D.ietf-hip-reload-instance] documents work in progress on using 3340 RELOAD with the HIP BONE. 3342 5.6.1.2. ICE-TCP 3344 The ICE-TCP draft [I-D.ietf-mmusic-ice-tcp] allows TCP to be 3345 supported as an Overlay Link protocol that can be added using ICE. 3347 5.6.1.3. Message-oriented Transports 3349 Modern message-oriented transports offer high performance, good 3350 congestion control, and avoid head of line blocking in case of lost 3351 data. These characteristics make them preferable as underlying 3352 transport protocols for RELOAD links. SCTP without message ordering 3353 and DCCP are two examples of such protocols. However, currently they 3354 are not well-supported by commonly available NATs, and specifications 3355 for ICE session establishment are not available. 3357 5.6.1.4. Tunneled Transports 3359 As of the time of this writing, there is significant interest in the 3360 IETF community in tunneling other transports over UDP, motivated by 3361 the situation that UDP is well-supported by modern NAT hardware, and 3362 similar performance can be achieved to native implementation. 3363 Currently SCTP, DCCP, and a generic tunneling extension are being 3364 proposed for message-oriented protocols. Once ICE traversal has been 3365 specified for these tunneled protocols, they should be 3366 straightforward to support as overlay link protocols. 3368 5.6.2. Framing Header 3370 In order to support unreliable links and to allow for quick detection 3371 of link failures when using reliable end-to-end transports, each 3372 message is wrapped in a very simple framing layer (FramedMessage) 3373 which is only used for each hop. This layer contains a sequence 3374 number which can then be used for ACKs. The same header is used for 3375 both reliable and unreliable transports for simplicity of 3376 implementation. 3378 The definition of FramedMessage is: 3380 enum { data(128), ack(129), (255)} FramedMessageType; 3382 struct { 3383 FramedMessageType type; 3385 select (type) { 3386 case data: 3387 uint32 sequence; 3388 opaque message<0..2^24-1>; 3390 case ack: 3391 uint32 ack_sequence; 3392 uint32 received; 3393 }; 3394 } FramedMessage; 3396 The type field of the PDU is set to indicate whether the message is 3397 data or an acknowledgement. 3399 If the message is of type "data", then the remainder of the PDU is as 3400 follows: 3402 sequence 3403 the sequence number. This increments by 1 for each framed message 3404 sent over this transport session. 3406 message 3407 the message that is being transmitted. 3409 Each connection has it own sequence number space. Initially the 3410 value is zero and it increments by exactly one for each message sent 3411 over that connection. 3413 When the receiver receives a message, it SHOULD immediately send an 3414 ACK message. The receiver MUST keep track of the 32 most recent 3415 sequence numbers received on this association in order to generate 3416 the appropriate ack. 3418 If the PDU is of type "ack", the contents are as follows: 3420 ack_sequence 3421 The sequence number of the message being acknowledged. 3423 received 3424 A bitmask indicating if each of the previous 32 sequence numbers 3425 before this packet has been among the 32 packets most recently 3426 received on this connection. When a packet is received with a 3427 sequence number N, the receiver looks at the sequence number of 3428 the previously 32 packets received on this connection. Call the 3429 previously received packet number M. For each of the previous 32 3430 packets, if the sequence number M is less than N but greater than 3431 N-32, the N-M bit of the received bitmask is set to one; otherwise 3432 it is zero. Note that a bit being set to one indicates positively 3433 that a particular packet was received, but a bit being set to zero 3434 means only that it is unknown whether or not the packet has been 3435 received, because it might have been received before the 32 most 3436 recently received packets. 3438 The received field bits in the ACK provide a high degree of 3439 redundancy so that the sender can figure out which packets the 3440 receiver has received and can then estimate packet loss rates. If 3441 the sender also keeps track of the time at which recent sequence 3442 numbers have been sent, the RTT can be estimated. 3444 5.6.3. Simple Reliability 3446 When RELOAD is carried over DTLS or another unreliable link protocol, 3447 it needs to be used with a reliability and congestion control 3448 mechanism, which is provided on a hop-by-hop basis. The basic 3449 principle is that each message, regardless of whether or not it 3450 carries a request or response, will get an ACK and be reliably 3451 retransmitted. The receiver's job is very simple, limited to just 3452 sending ACKs. All the complexity is at the sender side. This allows 3453 the sending implementation to trade off performance versus 3454 implementation complexity without affecting the wire protocol. 3456 5.6.3.1. Retransmission and Flow Control 3458 Because the receiver's role is limited to providing packet 3459 acknowledgements, a wide variety of congestion control algorithms can 3460 be implemented on the sender side while using the same basic wire 3461 protocol. In general, senders MAY implement any rate control scheme 3462 of their choice, provided that it is REQUIRED to be no more 3463 aggressive then TFRC[RFC5348]. 3465 The following section describes a simple, inefficient scheme that 3466 complies with this requirement. Another alternative would be TFRC-SP 3467 [RFC4828] and use the received bitmask to allow the sender to compute 3468 packet loss event rates. 3470 5.6.3.1.1. Trivial Retransmission 3472 A node SHOULD retransmit a message if it has not received an ACK 3473 after an interval of RTO ("Retransmission TimeOut"). The node MUST 3474 double the time to wait after each retransmission. In each 3475 retransmission, the sequence number is incremented. 3477 The RTO is an estimate of the round-trip time (RTT). Implementations 3478 can use a static value for RTO or a dynamic estimate which will 3479 result in better performance. For implementations that use a static 3480 value, the default value for RTO is 500 ms. Nodes MAY use smaller 3481 values of RTO if it is known that all nodes are within the local 3482 network. The default RTO MAY be chosen larger, and this is 3483 RECOMMENDED if it is known in advance (such as on high latency access 3484 links) that the round-trip time is larger. 3486 Implementations that use a dynamic estimate to compute the RTO MUST 3487 use the algorithm described in RFC 6298[RFC6298], with the exception 3488 that the value of RTO SHOULD NOT be rounded up to the nearest second 3489 but instead rounded up to the nearest millisecond. The RTT of a 3490 successful STUN transaction from the ICE stage is used as the initial 3491 measurement for formula 2.2 of RFC 6298. The sender keeps track of 3492 the time each message was sent for all recently sent messages. Any 3493 time an ACK is received, the sender can compute the RTT for that 3494 message by looking at the time the ACK was received and the time when 3495 the message was sent. This is used as a subsequent RTT measurement 3496 for formula 2.3 of RFC 6298 to update the RTO estimate. (Note that 3497 because retransmissions receive new sequence numbers, all received 3498 ACKs are used.) 3500 The value for RTO is calculated separately for each DTLS session. 3502 Retransmissions continue until a response is received, or until a 3503 total of 5 requests have been sent or there has been a hard ICMP 3504 error [RFC1122] or a TLS alert. The sender knows a response was 3505 received when it receives an ACK with a sequence number that 3506 indicates it is a response to one of the transmissions of this 3507 messages. For example, assuming an RTO of 500 ms, requests would be 3508 sent at times 0 ms, 500 ms, 1500 ms, 3500 ms, and 7500 ms. If all 3509 retransmissions for a message fail, then the sending node SHOULD 3510 close the connection routing the message. 3512 To determine when a link may be failing without waiting for the final 3513 timeout, observe when no ACKs have been received for an entire RTO 3514 interval, and then wait for three retransmissions to occur beyond 3515 that point. If no ACKs have been received by the time the third 3516 retransmission occurs, it is RECOMMENDED that the link be removed 3517 from the routing table. The link MAY be restored to the routing 3518 table if ACKs resume before the connection is closed, as described 3519 above. 3521 Once an ACK has been received for a message, the next message can be 3522 sent, but the peer SHOULD ensure that there is at least 10 ms between 3523 sending any two messages. The only time a value less than 10 ms can 3524 be used is when it is known that all nodes are on a network that can 3525 support retransmissions faster than 10 ms with no congestion issues. 3527 5.6.4. DTLS/UDP with SR 3529 This overlay link protocol consists of DTLS over UDP while 3530 implementing the Simple Reliability protocol. STUN Connectivity 3531 checks and keepalives are used. 3533 5.6.5. TLS/TCP with FH, No-ICE 3535 This overlay link protocol consists of TLS over TCP with the framing 3536 header. Because ICE is not used, STUN connectivity checks are not 3537 used upon establishing the TCP connection, nor are they used for 3538 keepalives. 3540 Because the TCP layer's application-level timeout is too slow to be 3541 useful for overlay routing, the Overlay Link implementation MUST use 3542 the framing header to measure the RTT of the connection and calculate 3543 an RTO as specified in Section 2 of [RFC6298]. The resulting RTO is 3544 not used for retransmissions, but as a timeout to indicate when the 3545 link SHOULD be removed from the routing table. It is RECOMMENDED 3546 that such a connection be retained for 30s to determine if the 3547 failure was transient before concluding the link has failed 3548 permanently. 3550 When sending candidates for TLS/TCP with FH, No-ICE, a passive 3551 candidate MUST be provided. 3553 5.6.6. DTLS/UDP with SR, No-ICE 3555 This overlay link protocol consists of DTLS over UDP while 3556 implementing the Simple Reliability protocol. Because ICE is not 3557 used, no STUN connectivity checks or keepalives are used. 3559 5.7. Fragmentation and Reassembly 3561 In order to allow transmission over datagram protocols such as DTLS, 3562 RELOAD messages may be fragmented. 3564 Any node along the path can fragment the message but only the final 3565 destination reassembles the fragments. When a node takes a packet 3566 and fragments it, each fragment has a full copy of the Forwarding 3567 Header but the data after the Forwarding Header is broken up in 3568 appropriate sized chunks. The size of the payload chunks needs to 3569 take into account space to allow the via and destination lists to 3570 grow. Each fragment MUST contain a full copy of the via and 3571 destination list and MUST contain at least 256 bytes of the message 3572 body. If the via and destination list are so large that this is not 3573 possible, RELOAD fragmentation is not performed and IP-layer 3574 fragmentation is allowed to occur. When a message must be 3575 fragmented, it SHOULD be split into equal-sized fragments that are no 3576 larger than the PMTU of the next overlay link minus 32 bytes. This 3577 is to allow the via list to grow before further fragmentation is 3578 required. 3580 Note that this fragmentation is not optimal for the end-to-end path - 3581 a message may be refragmented multiple times as it traverses the 3582 overlay but is only assembled at the final destination. This option 3583 has been chosen as it is far easier to implement than e2e PMTU 3584 discovery across an ever-changing overlay, and it effectively 3585 addresses the reliability issues of relying on IP-layer 3586 fragmentation. However, PING can be used to allow e2e PMTU to be 3587 implemented if desired. 3589 Upon receipt of a fragmented message by the intended peer, the peer 3590 holds the fragments in a holding buffer until the entire message has 3591 been received. The message is then reassembled into a single message 3592 and processed. In order to mitigate denial of service attacks, 3593 receivers SHOULD time out incomplete fragments after maximum request 3594 lifetime (15 seconds). Note this time was derived from looking at 3595 the end to end retransmission time and saving fragments long enough 3596 for the full end to end retransmissions to take place. Ideally the 3597 receiver would have enough buffer space to deal with as many 3598 fragments as can arrive in the maximum request lifetime. However, if 3599 the receiver runs out of buffer space to reassemble the messages it 3600 MUST drop the message. 3602 When a message is fragmented, the fragment offset value is stored in 3603 the lower 24 bits of the fragment field of the forwarding header. 3604 The offset is the number of bytes between the end of the forwarding 3605 header and the start of the data. The first fragment therefore has 3606 an offset of 0. The first and last bit indicators MUST be 3607 appropriately set. If the message is not fragmented, then both the 3608 first and last fragment bits are set to 1 and the offset is 0 3609 resulting in a fragment value of 0xC0000000. Note that this means 3610 that the first fragment bit is always 1, so isn't actually that 3611 useful. 3613 6. Data Storage Protocol 3615 RELOAD provides a set of generic mechanisms for storing and 3616 retrieving data in the Overlay Instance. These mechanisms can be 3617 used for new applications simply by defining new code points and a 3618 small set of rules. No new protocol mechanisms are required. 3620 The basic unit of stored data is a single StoredData structure: 3622 struct { 3623 uint32 length; 3624 uint64 storage_time; 3625 uint32 lifetime; 3626 StoredDataValue value; 3627 Signature signature; 3628 } StoredData; 3630 The contents of this structure are as follows: 3632 length 3633 The size of the StoredData structure in octets excluding the size 3634 of length itself. 3636 storage_time 3637 The time when the data was stored represented as the number of 3638 milliseconds elapsed since midnight Jan 1, 1970 UTC not counting 3639 leap seconds. This will have the same values for seconds as 3640 standard UNIX time or POSIX time. More information can be found 3641 at [UnixTime]. Any attempt to store a data value with a storage 3642 time before that of a value already stored at this location MUST 3643 generate a Error_Data_Too_Old error. This prevents rollback 3644 attacks. The node SHOULD make a best-effort attempt to use a 3645 correct clock to determine this number, however, the protocol does 3646 not require synchronized clocks: the receiving peer uses the 3647 storage time in the previous store, not its own clock. Clock 3648 values are used so that when clocks are generally synchronized, 3649 data may be stored in a single transaction, rather than querying 3650 for the value of a counter before the actual store. 3651 If a node attempting to store new data in response to a user 3652 request (rather than as an overlay maintenance operation such as 3653 occurs during unpartitioning) is rejected with an 3654 Error_Data_Too_Old error, the node MAY elect to perform its store 3655 using a storage_time that increments the value used with the 3656 previous store. This situation may occur when the clocks of nodes 3657 storing to this location are not properly synchronized. 3659 lifetime 3660 The validity period for the data, in seconds, starting from the 3661 time the peer receives the StoreReq. 3663 value 3664 The data value itself, as described in Section 6.2. 3666 signature 3667 A signature as defined in Section 6.1. 3669 Each Resource-ID specifies a single location in the Overlay Instance. 3670 However, each location may contain multiple StoredData values 3671 distinguished by Kind-ID. The definition of a kind describes both 3672 the data values which may be stored and the data model of the data. 3673 Some data models allow multiple values to be stored under the same 3674 Kind-ID. Section Section 6.2 describes the available data models. 3675 Thus, for instance, a given Resource-ID might contain a single-value 3676 element stored under Kind-ID X and an array containing multiple 3677 values stored under Kind-ID Y. 3679 6.1. Data Signature Computation 3681 Each StoredData element is individually signed. However, the 3682 signature also must be self-contained and cover the Kind-ID and 3683 Resource-ID even though they are not present in the StoredData 3684 structure. The input to the signature algorithm is: 3686 resource_id || kind || storage_time || StoredDataValue || 3687 SignerIdentity 3689 Where || indicates concatenation. 3691 Where these values are: 3693 resource_id 3694 The resource ID where this data is stored. 3696 kind 3697 The Kind-ID for this data. 3699 storage_time 3701 The contents of the storage_time data value. 3702 StoredDataValue 3703 The contents of the stored data value, as described in the 3704 previous sections. 3706 SignerIdentity 3707 The signer identity as defined in Section 5.3.4. 3709 Once the signature has been computed, the signature is represented 3710 using a signature element, as described in Section 5.3.4. 3712 6.2. Data Models 3714 The protocol currently defines the following data models: 3716 o single value 3717 o array 3718 o dictionary 3720 These are represented with the StoredDataValue structure. The actual 3721 dataModel is known from the kind being stored. 3723 struct { 3724 Boolean exists; 3725 opaque value<0..2^32-1>; 3726 } DataValue; 3728 struct { 3729 select (dataModel) { 3730 case single_value: 3731 DataValue single_value_entry; 3733 case array: 3734 ArrayEntry array_entry; 3736 case dictionary: 3737 DictionaryEntry dictionary_entry; 3739 /* This structure may be extended */ 3740 }; 3741 } StoredDataValue; 3743 We now discuss the properties of each data model in turn: 3745 6.2.1. Single Value 3747 A single-value element is a simple sequence of bytes. There may be 3748 only one single-value element for each Resource-ID, Kind-ID pair. 3750 A single value element is represented as a DataValue, which contains 3751 the following two elements: 3753 exists 3754 This value indicates whether the value exists at all. If it is 3755 set to False, it means that no value is present. If it is True, 3756 that means that a value is present. This gives the protocol a 3757 mechanism for indicating nonexistence as opposed to emptiness. 3759 value 3760 The stored data. 3762 6.2.2. Array 3764 An array is a set of opaque values addressed by an integer index. 3765 Arrays are zero based. Note that arrays can be sparse. For 3766 instance, a Store of "X" at index 2 in an empty array produces an 3767 array with the values [ NA, NA, "X"]. Future attempts to fetch 3768 elements at index 0 or 1 will return values with "exists" set to 3769 False. 3771 A array element is represented as an ArrayEntry: 3773 struct { 3774 uint32 index; 3775 DataValue value; 3776 } ArrayEntry; 3778 The contents of this structure are: 3780 index 3781 The index of the data element in the array. 3783 value 3784 The stored data. 3786 6.2.3. Dictionary 3788 A dictionary is a set of opaque values indexed by an opaque key with 3789 one value for each key. A single dictionary entry is represented as 3790 follows: 3792 A dictionary element is represented as a DictionaryEntry: 3794 typedef opaque DictionaryKey<0..2^16-1>; 3796 struct { 3797 DictionaryKey key; 3798 DataValue value; 3799 } DictionaryEntry; 3801 The contents of this structure are: 3803 key 3804 The dictionary key for this value. 3806 value 3807 The stored data. 3809 6.3. Access Control Policies 3811 Every kind which is storable in an overlay MUST be associated with an 3812 access control policy. This policy defines whether a request from a 3813 given node to operate on a given value should succeed or fail. It is 3814 anticipated that only a small number of generic access control 3815 policies are required. To that end, this section describes a small 3816 set of such policies and Section 13.4 establishes a registry for new 3817 policies if required. Each policy has a short string identifier 3818 which is used to reference it in the configuration document. 3820 In the following policies, the term "signer" refers to the signer of 3821 the StoredValue object and, in the case of non-replica stores, to the 3822 signer of the StoreReq message. I.e., in a non-replica store, both 3823 the signer of the StoredValue and the signer of the StoreReq MUST 3824 conform to the policy. In the case of a replica store, the signer of 3825 the StoredValue MUST conform to the policy and the StoreReq itself 3826 MUST be checked as described in Section 6.4.1.1. 3828 6.3.1. USER-MATCH 3830 In the USER-MATCH policy, a given value MUST be written (or 3831 overwritten) if and only if the signer's certificate has a user name 3832 which hashes (using the hash function for the overlay) to the 3833 Resource-ID for the resource. Recall that the certificate may, 3834 depending on the overlay configuration, be self-signed. 3836 6.3.2. NODE-MATCH 3838 In the NODE-MATCH policy, a given value MUST be written (or 3839 overwritten) if and only if the signer's certificate has a specified 3840 Node-ID which hashes (using the hash function for the overlay) to the 3841 Resource-ID for the resource and that Node-ID is the one indicated in 3842 the SignerIdentity value cert_hash. 3844 6.3.3. USER-NODE-MATCH 3846 The USER-NODE-MATCH policy may only be used with dictionary types. 3847 In the USER-NODE-MATCH policy, a given value MUST be written (or 3848 overwritten) if and only if the signer's certificate has a user name 3849 which hashes (using the hash function for the overlay) to the 3850 Resource-ID for the resource. In addition, the dictionary key MUST 3851 be equal to the Node-ID in the certificate and that Node-ID MUST be 3852 the one indicated in the SignerIdentity value cert_hash. 3854 6.3.4. NODE-MULTIPLE 3856 In the NODE-MULTIPLE policy, a given value MUST be written (or 3857 overwritten) if and only if signer's certificate contains a Node-ID 3858 such that H(Node-ID || i) is equal to the Resource-ID for some small 3859 integer value of i and that Node-ID is the one indicated in the 3860 SignerIdentity value cert_hash. When this policy is in use, the 3861 maximum value of i MUST be specified in the kind definition. 3863 Note that as i is not carried on the wire, the verifier MUST iterate 3864 through potential i values up to the maximum value in order to 3865 determine whether a store is acceptable. 3867 6.4. Data Storage Methods 3869 RELOAD provides several methods for storing and retrieving data: 3871 o Store values in the overlay 3872 o Fetch values from the overlay 3873 o Stat: get metadata about values in the overlay 3874 o Find the values stored at an individual peer 3876 These methods are each described in the following sections. 3878 6.4.1. Store 3880 The Store method is used to store data in the overlay. The format of 3881 the Store request depends on the data model which is determined by 3882 the kind. 3884 6.4.1.1. Request Definition 3886 A StoreReq message is a sequence of StoreKindData values, each of 3887 which represents a sequence of stored values for a given kind. The 3888 same Kind-ID MUST NOT be used twice in a given store request. Each 3889 value is then processed in turn. These operations MUST be atomic. 3890 If any operation fails, the state MUST be rolled back to before the 3891 request was received. 3893 The store request is defined by the StoreReq structure: 3895 struct { 3896 KindId kind; 3897 uint64 generation_counter; 3898 StoredData values<0..2^32-1>; 3899 } StoreKindData; 3901 struct { 3902 ResourceId resource; 3903 uint8 replica_number; 3904 StoreKindData kind_data<0..2^32-1>; 3905 } StoreReq; 3907 A single Store request stores data of a number of kinds to a single 3908 resource location. The contents of the structure are: 3910 resource 3911 The resource to store at. 3913 replica_number 3914 The number of this replica. When a storing peer saves replicas to 3915 other peers each peer is assigned a replica number starting from 1 3916 and sent in the Store message. This field is set to 0 when a node 3917 is storing its own data. This allows peers to distinguish replica 3918 writes from original writes. 3920 kind_data 3921 A series of elements, one for each kind of data to be stored. 3923 If the replica number is zero, then the peer MUST check that it is 3924 responsible for the resource and, if not, reject the request. If the 3925 replica number is nonzero, then the peer MUST check that it expects 3926 to be a replica for the resource and that the request sender is 3927 consistent with being the responsible node (i.e., that the receiving 3928 peer does not know of a better node) and, if not, reject the request. 3930 Each StoreKindData element represents the data to be stored for a 3931 single Kind-ID. The contents of the element are: 3933 kind 3934 The Kind-ID. Implementations MUST reject requests corresponding 3935 to unknown kinds. 3937 generation_counter 3938 The expected current state of the generation counter 3939 (approximately the number of times this object has been written; 3940 see below for details). 3942 values 3943 The value or values to be stored. This may contain one or more 3944 stored_data values depending on the data model associated with 3945 each kind. 3947 The peer MUST perform the following checks: 3949 o The Kind-ID is known and supported. 3950 o The signatures over each individual data element (if any) are 3951 valid. If this check fails, the request MUST be rejected with an 3952 Error_Forbidden error. 3953 o Each element is signed by a credential which is authorized to 3954 write this kind at this Resource-ID. If this check fails, the 3955 request MUST be rejected with an Error_Forbidden error. 3957 o For original (non-replica) stores, the StoreReq is signed by a 3958 credential which is authorized to write this kind at this 3959 Resource-Id. If this check fails, the request MUST be rejected 3960 with an Error_Forbidden error. 3961 o For replica stores, the StoreReq is signed by a Node-Id which is a 3962 plausible node to either have originally stored the value or in 3963 the replica set. What this means is overlay specific, but in the 3964 case of the Chord based DHT defined in this specification, replica 3965 StoreReqs MUST come from nodes which are either in the known 3966 replica set for a given resource or which are closer than some 3967 node in the replica set. If this check fails, the request MUST be 3968 rejected with an Error_Forbidden error. 3969 o For original (non-replica) stores, the peer MUST check that if the 3970 generation counter is non-zero, it equals the current value of the 3971 generation counter for this kind. This feature allows the 3972 generation counter to be used in a way similar to the HTTP Etag 3973 feature. 3974 o For replica Stores, the peer MUST set the generation counter to 3975 match the generation counter in the message, and MUST NOT check 3976 the generation counter against the current value. Replica Stores 3977 MUST NOT use a generation counter of 0. 3978 o The storage time values are greater than that of any value which 3979 would be replaced by this Store. 3980 o The size and number of the stored values is consistent with the 3981 limits specified in the overlay configuration. 3982 o If the data is signed with identity_type set to "none" and/or 3983 SignatureAndHashAlgorithm values set to {0, 0} ("anonymous" and 3984 "none"), the StoreReq MUST be rejected with an Error_forbidden 3985 error. Only synthesized data returned by the storage can use 3986 these values 3988 If all these checks succeed, the peer MUST attempt to store the data 3989 values. For non-replica stores, if the store succeeds and the data 3990 is changed, then the peer must increase the generation counter by at 3991 least one. If there are multiple stored values in a single 3992 StoreKindData, it is permissible for the peer to increase the 3993 generation counter by only 1 for the entire Kind-ID, or by 1 or more 3994 than one for each value. Accordingly, all stored data values must 3995 have a generation counter of 1 or greater. 0 is used in the Store 3996 request to indicate that the generation counter should be ignored for 3997 processing this request; however the responsible peer should increase 3998 the stored generation counter and should return the correct 3999 generation counter in the response. 4001 When a peer stores data previously stored by another node (e.g., for 4002 replicas or topology shifts) it MUST adjust the lifetime value 4003 downward to reflect the amount of time the value was stored at the 4004 peer. The adjustment SHOULD be implemented by an algorithm 4005 equivalent to the following: at the time the peer initially receives 4006 the StoreReq it notes the local time T. When it then attempts to do a 4007 StoreReq to another node it should decrement the lifetime value by 4008 the difference between the current local time and T. 4010 Unless otherwise specified by the usage, if a peer attempts to store 4011 data previously stored by another node (e.g., for replicas or 4012 topology shifts) and that store fails with either an 4013 Error_Generation_Counter_Too_Low or an Error_Data_Too old error, the 4014 peer MUST fetch the newer data from the peer generating the error and 4015 use that to replace its own copy. This rule allows resynchronization 4016 after partitions heal. 4018 The properties of stores for each data model are as follows: 4020 Single-value: 4021 A store of a new single-value element creates the element if it 4022 does not exist and overwrites any existing value with the new 4023 value. 4025 Array: 4026 A store of an array entry replaces (or inserts) the given value at 4027 the location specified by the index. Because arrays are sparse, a 4028 store past the end of the array extends it with nonexistent values 4029 (exists=False) as required. A store at index 0xffffffff places 4030 the new value at the end of the array regardless of the length of 4031 the array. The resulting StoredData has the correct index value 4032 when it is subsequently fetched. 4034 Dictionary: 4035 A store of a dictionary entry replaces (or inserts) the given 4036 value at the location specified by the dictionary key. 4038 The following figure shows the relationship between these structures 4039 for an example store which stores the following values at resource 4040 "1234" 4042 o The value "abc" in the single value location for kind X 4043 o The value "foo" at index 0 in the array for kind Y 4044 o The value "bar" at index 1 in the array for kind Y 4045 Store 4046 resource=1234 4047 replica_number = 0 4048 / \ 4049 / \ 4050 StoreKindData StoreKindData 4051 kind=X (Single-Value) kind=Y (Array) 4052 generation_counter = 99 generation_counter = 107 4053 | /\ 4054 | / \ 4055 StoredData / \ 4056 storage_time = xxxxxxx / \ 4057 lifetime = 86400 / \ 4058 signature = XXXX / \ 4059 | | | 4060 | StoredData StoredData 4061 | storage_time = storage_time = 4062 | yyyyyyyy zzzzzzz 4063 | lifetime = 86400 lifetime = 33200 4064 | signature = YYYY signature = ZZZZ 4065 | | | 4066 StoredDataValue | | 4067 value="abc" | | 4068 | | 4069 StoredDataValue StoredDataValue 4070 index=0 index=1 4071 value="foo" value="bar" 4073 6.4.1.2. Response Definition 4075 In response to a successful Store request the peer MUST return a 4076 StoreAns message containing a series of StoreKindResponse elements 4077 containing the current value of the generation counter for each 4078 Kind-ID, as well as a list of the peers where the data will be 4079 replicated by the node processing the request. 4081 struct { 4082 KindId kind; 4083 uint64 generation_counter; 4084 NodeId replicas<0..2^16-1>; 4085 } StoreKindResponse; 4087 struct { 4088 StoreKindResponse kind_responses<0..2^16-1>; 4089 } StoreAns; 4091 The contents of each StoreKindResponse are: 4093 kind 4094 The Kind-ID being represented. 4096 generation_counter 4097 The current value of the generation counter for that Kind-ID. 4099 replicas 4100 The list of other peers at which the data was/will be replicated. 4101 In overlays and applications where the responsible peer is 4102 intended to store redundant copies, this allows the storing peer 4103 to independently verify that the replicas have in fact been 4104 stored. It does this verification by using the Stat method (see 4105 Section 6.4.3). Note that the storing peer is not required to 4106 perform this verification. 4108 The response itself is just StoreKindResponse values packed end-to- 4109 end. 4111 If any of the generation counters in the request precede the 4112 corresponding stored generation counter, then the peer MUST fail the 4113 entire request and respond with an Error_Generation_Counter_Too_Low 4114 error. The error_info in the ErrorResponse MUST be a StoreAns 4115 response containing the correct generation counter for each kind and 4116 the replica list, which will be empty. For original (non-replica) 4117 stores, a node which receives such an error SHOULD attempt to fetch 4118 the data and, if the storage_time value is newer, replace its own 4119 data with that newer data. This rule improves data consistency in 4120 the case of partitions and merges. 4122 If the data being stored is too large for the allowed limit by the 4123 given usage, then the peer MUST fail the request and generate an 4124 Error_Data_Too_Large error. 4126 If any type of request tries to access a data kind that the node does 4127 not know about, an Error_Unknown_Kind MUST be generated. The 4128 error_info in the Error_Response is: 4130 KindId unknown_kinds<0..2^8-1>; 4132 which lists all the kinds that were unrecognized. A node which 4133 receives this error MUST generate a ConfigUpdate message which 4134 contains the appropriate kind definition (assuming that in fact a 4135 kind was used which was defined in the configuration document). 4137 6.4.1.3. Removing Values 4139 This version of RELOAD (unlike previous versions) does not have an 4140 explicit Remove operation. Rather, values are Removed by storing 4141 "nonexistent" values in their place. Each DataValue contains a 4142 boolean value called "exists" which indicates whether a value is 4143 present at that location. In order to effectively remove a value, 4144 the owner stores a new DataValue with "exists" set to "false": 4146 exists = false 4147 value = {} (0 length) 4149 The owner SHOULD use a lifetime for the nonexistent value at least as 4150 long as the remainder of the lifetime of the value it is replacing; 4151 otherwise it is possible for the original value to be accidentally or 4152 maliciously re-stored after the storing node has expired it. Note 4153 that there is still a window of vulnerability for replay attack after 4154 the original lifetime has expired (as with any store). This attack 4155 can be mitigated by doing a nonexistent store with a very long 4156 lifetime. 4158 Storing nodes MUST treat these nonexistent values the same way they 4159 treat any other stored value, including overwriting the existing 4160 value, replicating them, and aging them out as necessary when 4161 lifetime expires. When a stored nonexistent value's lifetime 4162 expires, it is simply removed from the storing node like any other 4163 stored value expiration. 4165 Note that in the case of arrays and dictionaries, expiration may 4166 create an implicit, unsigned "nonexistent" value to represent a gap 4167 in the data structure, as might happen when any value is aged out. 4168 However, this value isn't persistent nor is it replicated. It is 4169 simply synthesized by the storing node. 4171 6.4.2. Fetch 4173 The Fetch request retrieves one or more data elements stored at a 4174 given Resource-ID. A single Fetch request can retrieve multiple 4175 different kinds. 4177 6.4.2.1. Request Definition 4179 struct { 4180 int32 first; 4181 int32 last; 4182 } ArrayRange; 4184 struct { 4185 KindId kind; 4186 uint64 generation; 4187 uint16 length; 4189 select (dataModel) { 4190 case single_value: ; /* Empty */ 4192 case array: 4193 ArrayRange indices<0..2^16-1>; 4195 case dictionary: 4196 DictionaryKey keys<0..2^16-1>; 4198 /* This structure may be extended */ 4200 } model_specifier; 4201 } StoredDataSpecifier; 4203 struct { 4204 ResourceId resource; 4205 StoredDataSpecifier specifiers<0..2^16-1>; 4206 } FetchReq; 4208 The contents of the Fetch requests are as follows: 4210 resource 4211 The Resource-ID to fetch from. 4213 specifiers 4214 A sequence of StoredDataSpecifier values, each specifying some of 4215 the data values to retrieve. 4217 Each StoredDataSpecifier specifies a single kind of data to retrieve 4218 and (if appropriate) the subset of values that are to be retrieved. 4219 The contents of the StoredDataSpecifier structure are as follows: 4221 kind 4222 The Kind-ID of the data being fetched. Implementations SHOULD 4223 reject requests corresponding to unknown kinds unless specifically 4224 configured otherwise. 4226 dataModel 4227 The data model of the data. This is not transmitted on the wire 4228 but comes from the definition of the kind. 4230 generation 4231 The last generation counter that the requesting node saw. This 4232 may be used to avoid unnecessary fetches or it may be set to zero. 4234 length 4235 The length of the rest of the structure, thus allowing 4236 extensibility. 4238 model_specifier 4239 A reference to the data value being requested within the data 4240 model specified for the kind. For instance, if the data model is 4241 "array", it might specify some subset of the values. 4243 The model_specifier is as follows: 4245 o If the data model is single value, the specifier is empty. 4246 o If the data model is array, the specifier contains a list of 4247 ArrayRange elements, each of which contains two integers. The 4248 first integer is the beginning of the range and the second is the 4249 end of the range. 0 is used to indicate the first element and 4250 0xffffffff is used to indicate the final element. The first 4251 integer must be less than the second. While multiple ranges MAY 4252 be specified, they MUST NOT overlap. 4253 o If the data model is dictionary then the specifier contains a list 4254 of the dictionary keys being requested. If no keys are specified, 4255 than this is a wildcard fetch and all key-value pairs are 4256 returned. 4258 The generation counter is used to indicate the requester's expected 4259 state of the storing peer. If the generation counter in the request 4260 matches the stored counter, then the storing peer returns a response 4261 with no StoredData values. 4263 Note that because the certificate for a user is typically stored at 4264 the same location as any data stored for that user, a requesting node 4265 that does not already have the user's certificate should request the 4266 certificate in the Fetch as an optimization. 4268 6.4.2.2. Response Definition 4270 The response to a successful Fetch request is a FetchAns message 4271 containing the data requested by the requester. 4273 struct { 4274 KindId kind; 4275 uint64 generation; 4276 StoredData values<0..2^32-1>; 4277 } FetchKindResponse; 4279 struct { 4280 FetchKindResponse kind_responses<0..2^32-1>; 4281 } FetchAns; 4283 The FetchAns structure contains a series of FetchKindResponse 4284 structures. There MUST be one FetchKindResponse element for each 4285 Kind-ID in the request. 4287 The contents of the FetchKindResponse structure are as follows: 4289 kind 4290 the kind that this structure is for. 4292 generation 4293 the generation counter for this kind. 4295 values 4296 the relevant values. If the generation counter in the request 4297 matches the generation counter in the stored data, then no 4298 StoredData values are returned. Otherwise, all relevant data 4299 values MUST be returned. A nonexistent value (i.e., one which the 4300 node has no knowledge of) is represented by a synthetic value with 4301 "exists" set to False and has an empty signature. Specifically, 4302 the identity_type is set to "none", the SignatureAndHashAlgorithm 4303 values are set to {0, 0} ("anonymous" and "none" respectively), 4304 and the signature value is of zero length. This removes the need 4305 for the responding node to do signatures for values which do not 4306 exist. These signatures are unnecessary as the entire response is 4307 signed by that node. Note that entries which have been removed by 4308 the procedure of Section 6.4.1.3 and have not yet expired also 4309 have exists = false but have valid signatures from the node which 4310 did the store. 4312 There is one subtle point about signature computation on arrays. If 4313 the storing node uses the append feature (where the 4314 index=0xffffffff), then the index in the StoredData that is returned 4315 will not match that used by the storing node, which would break the 4316 signature. In order to avoid this issue, the index value in the 4317 array is set to zero before the signature is computed. This implies 4318 that malicious storing nodes can reorder array entries without being 4319 detected. 4321 6.4.3. Stat 4323 The Stat request is used to get metadata (length, generation counter, 4324 digest, etc.) for a stored element without retrieving the element 4325 itself. The name is from the UNIX stat(2) system call which performs 4326 a similar function for files in a file system. It also allows the 4327 requesting node to get a list of matching elements without requesting 4328 the entire element. 4330 6.4.3.1. Request Definition 4332 The Stat request is identical to the Fetch request. It simply 4333 specifies the elements to get metadata about. 4335 struct { 4336 ResourceId resource; 4337 StoredDataSpecifier specifiers<0..2^16-1>; 4338 } StatReq; 4340 6.4.3.2. Response Definition 4342 The Stat response contains the same sort of entries that a Fetch 4343 response would contain; however, instead of containing the element 4344 data it contains metadata. 4346 struct { 4347 Boolean exists; 4348 uint32 value_length; 4349 HashAlgorithm hash_algorithm; 4350 opaque hash_value<0..255>; 4351 } MetaData; 4353 struct { 4354 uint32 index; 4355 MetaData value; 4356 } ArrayEntryMeta; 4358 struct { 4359 DictionaryKey key; 4360 MetaData value; 4361 } DictionaryEntryMeta; 4363 struct { 4364 select (model) { 4365 case single_value: 4366 MetaData single_value_entry; 4368 case array: 4369 ArrayEntryMeta array_entry; 4371 case dictionary: 4372 DictionaryEntryMeta dictionary_entry; 4374 /* This structure may be extended */ 4375 }; 4376 } MetaDataValue; 4378 struct { 4379 uint32 value_length; 4380 uint64 storage_time; 4381 uint32 lifetime; 4382 MetaDataValue metadata; 4383 } StoredMetaData; 4385 struct { 4386 KindId kind; 4387 uint64 generation; 4388 StoredMetaData values<0..2^32-1>; 4389 } StatKindResponse; 4391 struct { 4392 StatKindResponse kind_responses<0..2^32-1>; 4393 } StatAns; 4395 The structures used in StatAns parallel those used in FetchAns: a 4396 response consists of multiple StatKindResponse values, one for each 4397 kind that was in the request. The contents of the StatKindResponse 4398 are the same as those in the FetchKindResponse, except that the 4399 values list contains StoredMetaData entries instead of StoredData 4400 entries. 4402 The contents of the StoredMetaData structure are the same as the 4403 corresponding fields in StoredData except that there is no signature 4404 field and the value is a MetaDataValue rather than a StoredDataValue. 4406 A MetaDataValue is a variant structure, like a StoredDataValue, 4407 except for the types of each arm, which replace DataValue with 4408 MetaData. 4410 The only really new structure is MetaData, which has the following 4411 contents: 4413 exists 4414 Same as in DataValue 4416 value_length 4417 The length of the stored value. 4419 hash_algorithm 4420 The hash algorithm used to perform the digest of the value. 4422 hash_value 4423 A digest of the value using hash_algorithm. 4425 6.4.4. Find 4427 The Find request can be used to explore the Overlay Instance. A Find 4428 request for a Resource-ID R and a Kind-ID T retrieves the Resource-ID 4429 (if any) of the resource of kind T known to the target peer which is 4430 closest to R. This method can be used to walk the Overlay Instance by 4431 iteratively fetching R_n+1=nearest(1 + R_n). 4433 6.4.4.1. Request Definition 4435 The FindReq message contains a Resource-ID and a series of Kind-IDs 4436 identifying the resource the peer is interested in. 4438 struct { 4439 ResourceId resource; 4440 KindId kinds<0..2^8-1>; 4441 } FindReq; 4443 The request contains a list of Kind-IDs which the Find is for, as 4444 indicated below: 4446 resource 4447 The desired Resource-ID 4449 kinds 4450 The desired Kind-IDs. Each value MUST only appear once, and if 4451 not the request MUST be rejected with an error. 4453 6.4.4.2. Response Definition 4455 A response to a successful Find request is a FindAns message 4456 containing the closest Resource-ID on the peer for each kind 4457 specified in the request. 4459 struct { 4460 KindId kind; 4461 ResourceId closest; 4462 } FindKindData; 4464 struct { 4465 FindKindData results<0..2^16-1>; 4466 } FindAns; 4468 If the processing peer is not responsible for the specified 4469 Resource-ID, it SHOULD return an Error_Not_Found error code. 4471 For each Kind-ID in the request the response MUST contain a 4472 FindKindData indicating the closest Resource-ID for that Kind-ID, 4473 unless the kind is not allowed to be used with Find in which case a 4474 FindKindData for that Kind-ID MUST NOT be included in the response. 4475 If a Kind-ID is not known, then the corresponding Resource-ID MUST be 4476 0. Note that different Kind-IDs may have different closest Resource- 4477 IDs. 4479 The response is simply a series of FindKindData elements, one per 4480 kind, concatenated end-to-end. The contents of each element are: 4482 kind 4483 The Kind-ID. 4485 closest 4486 The closest resource ID to the specified resource ID. This is 0 4487 if no resource ID is known. 4489 Note that the response does not contain the contents of the data 4490 stored at these Resource-IDs. If the requester wants this, it must 4491 retrieve it using Fetch. 4493 6.4.5. Defining New Kinds 4495 There are two ways to define a new kind. The first is by writing a 4496 document and registering the kind-id with IANA. This is the 4497 preferred method for kinds which may be widely used and reused. The 4498 second method is to simply define the kind and its parameters in the 4499 configuration document using the section of kind-id space set aside 4500 for private use. This method MAY be used to define ad hoc kinds in 4501 new overlays. 4503 However a kind is defined, the definition must include: 4505 o The meaning of the data to be stored (in some textual form). 4506 o The Kind-ID. 4507 o The data model (single value, array, dictionary, etc). 4508 o The access control model. 4510 In addition, when kinds are registered with IANA, each kind is 4511 assigned a short string name which is used to refer to it in 4512 configuration documents. 4514 While each kind needs to define what data model is used for its data, 4515 that does not mean that it must define new data models. Where 4516 practical, kinds should use the existing data models. The intention 4517 is that the basic data model set be sufficient for most applications/ 4518 usages. 4520 7. Certificate Store Usage 4522 The Certificate Store usage allows a peer to store its certificate in 4523 the overlay, thus avoiding the need to send a certificate in each 4524 message - a reference may be sent instead. 4526 A user/peer MUST store its certificate at Resource-IDs derived from 4527 two Resource Names: 4529 o The user name in the certificate. 4531 o The Node-ID in the certificate. 4533 Note that in the second case the certificate is not stored at the 4534 peer's Node-ID but rather at a hash of the peer's Node-ID. The 4535 intention here (as is common throughout RELOAD) is to avoid making a 4536 peer responsible for its own data. 4538 A peer MUST ensure that the user's certificates are stored in the 4539 Overlay Instance. New certificates are stored at the end of the 4540 list. This structure allows users to store an old and a new 4541 certificate that both have the same Node-ID, which allows for 4542 migration of certificates when they are renewed. 4544 This usage defines the following kinds: 4546 Name: CERTIFICATE_BY_NODE 4548 Data Model: The data model for CERTIFICATE_BY_NODE data is array. 4550 Access Control: NODE-MATCH. 4552 Name: CERTIFICATE_BY_USER 4554 Data Model: The data model for CERTIFICATE_BY_USER data is array. 4556 Access Control: USER-MATCH. 4558 8. TURN Server Usage 4560 The TURN server usage allows a RELOAD peer to advertise that it is 4561 prepared to be a TURN server as defined in [RFC5766]. When a node 4562 starts up, it joins the overlay network and forms several connections 4563 in the process. If the ICE stage in any of these connections returns 4564 a reflexive address that is not the same as the peer's perceived 4565 address, then the peer is behind a NAT and not a candidate for a TURN 4566 server. Additionally, if the peer's IP address is in the private 4567 address space range, then it is also not a candidate for a TURN 4568 server. Otherwise, the peer SHOULD assume it is a potential TURN 4569 server and follow the procedures below. 4571 If the node is a candidate for a TURN server it will insert some 4572 pointers in the overlay so that other peers can find it. The overlay 4573 configuration file specifies a turn-density parameter that indicates 4574 how many times each TURN server should record itself in the overlay. 4575 Typically this should be set to the reciprocal of the estimate of 4576 what percentage of peers will act as TURN servers. If the turn- 4577 density is not set to zero, for each value, called d, between 1 and 4578 turn-density, the peer forms a Resource Name by concatenating its 4579 Node-ID and the value d. This Resource Name is hashed to form a 4580 Resource-ID. The address of the peer is stored at that Resource-ID 4581 using type TURN-SERVICE and the TurnServer object: 4583 struct { 4584 uint8 iteration; 4585 IpAddressAndPort server_address; 4586 } TurnServer; 4588 The contents of this structure are as follows: 4590 iteration 4591 the d value 4593 server_address 4594 the address at which the TURN server can be contacted. 4596 Note: Correct functioning of this algorithm depends on having turn- 4597 density be an reasonable estimate of the reciprocal of the 4598 proportion of nodes in the overlay that can act as TURN servers. 4599 If the turn-density value in the configuration file is too low, 4600 then the process of finding TURN servers becomes more expensive as 4601 multiple candidate Resource-IDs must be probed to find a TURN 4602 server. 4604 Peers that provide this service need to support the TURN extensions 4605 to STUN for media relay as defined in [RFC5766]. 4607 This usage defines the following kind to indicate that a peer is 4608 willing to act as a TURN server: 4610 Name TURN-SERVICE 4611 Data Model The TURN-SERVICE kind stores a single value for each 4612 Resource-ID. 4613 Access Control NODE-MULTIPLE, with maximum iteration counter 20. 4615 Peers can find other servers by selecting a random Resource-ID and 4616 then doing a Find request for the appropriate Kind-ID with that 4617 Resource-ID. The Find request gets routed to a random peer based on 4618 the Resource-ID. If that peer knows of any servers, they will be 4619 returned. The returned response may be empty if the peer does not 4620 know of any servers, in which case the process gets repeated with 4621 some other random Resource-ID. As long as the ratio of servers 4622 relative to peers is not too low, this approach will result in 4623 finding a server relatively quickly. 4625 9. Chord Algorithm 4627 This algorithm is assigned the name chord-reload to indicate it is an 4628 adaptation of the basic Chord based DHT algorithm. 4630 This algorithm differs from the originally presented Chord algorithm 4631 [Chord]. It has been updated based on more recent research results 4632 and implementation experiences, and to adapt it to the RELOAD 4633 protocol. A short list of differences: 4635 o The original Chord algorithm specified that a single predecessor 4636 and a successor list be stored. The chord-reload algorithm 4637 attempts to have more than one predecessor and successor. The 4638 predecessor sets help other neighbors learn their successor list. 4639 o The original Chord specification and analysis called for iterative 4640 routing. RELOAD specifies recursive routing. In addition to the 4641 performance implications, the cost of NAT traversal dictates 4642 recursive routing. 4643 o Finger table entries are indexed in opposite order. Original 4644 Chord specifies finger[0] as the immediate successor of the peer. 4645 chord-reload specifies finger[0] as the peer 180 degrees around 4646 the ring from the peer. This change was made to simplify 4647 discussion and implementation of variable sized finger tables. 4648 However, with either approach no more than O(log N) entries should 4649 typically be stored in a finger table. 4650 o The stabilize() and fix_fingers() algorithms in the original Chord 4651 algorithm are merged into a single periodic process. 4652 Stabilization is implemented slightly differently because of the 4653 larger neighborhood, and fix_fingers is not as aggressive to 4654 reduce load, nor does it search for optimal matches of the finger 4655 table entries. 4656 o RELOAD uses a 128 bit hash instead of a 160 bit hash, as RELOAD is 4657 not designed to be used in networks with close to or more than 4658 2^128 nodes (and it is hard to see how one would assemble such a 4659 network). 4660 o RELOAD uses randomized finger entries as described in 4661 Section 9.7.4.2. 4662 o This algorithm allows the use of either reactive or periodic 4663 recovery. The original Chord paper used periodic recovery. 4664 Reactive recovery provides better performance in small overlays, 4665 but is believed to be unstable in large (>1000) overlays with high 4666 levels of churn [handling-churn-usenix04]. The overlay 4667 configuration file specifies a "chord-reactive" element that 4668 indicates whether reactive recovery should be used. 4670 9.1. Overview 4672 The algorithm described here is a modified version of the Chord 4673 algorithm. Each peer keeps track of a finger table and a neighbor 4674 table. The neighbor table contains at least the three peers before 4675 and after this peer in the DHT ring. There may not be three entries 4676 in all cases such as small rings or while the ring topology is 4677 changing. The first entry in the finger table contains the peer 4678 half-way around the ring from this peer; the second entry contains 4679 the peer that is 1/4 of the way around; the third entry contains the 4680 peer that is 1/8th of the way around, and so on. Fundamentally, the 4681 chord data structure can be thought of a doubly-linked list formed by 4682 knowing the successors and predecessor peers in the neighbor table, 4683 sorted by the Node-ID. As long as the successor peers are correct, 4684 the DHT will return the correct result. The pointers to the prior 4685 peers are kept to enable the insertion of new peers into the list 4686 structure. Keeping multiple predecessor and successor pointers makes 4687 it possible to maintain the integrity of the data structure even when 4688 consecutive peers simultaneously fail. The finger table forms a skip 4689 list, so that entries in the linked list can be found in O(log(N)) 4690 time instead of the typical O(N) time that a linked list would 4691 provide. 4693 A peer, n, is responsible for a particular Resource-ID k if k is less 4694 than or equal to n and k is greater than p, where p is the Node-ID of 4695 the previous peer in the neighbor table. Care must be taken when 4696 computing to note that all math is modulo 2^128. 4698 9.2. Hash Function 4700 For this Chord based topology plugin, the size of the Resource-ID is 4701 128 bits. The hash of a Resource-ID is computed using SHA-1 4702 [RFC3174]then truncating the SHA-1 result to the most significant 128 4703 bits. 4705 9.3. Routing 4707 The routing table is the union of the neighbor table and the finger 4708 table. 4710 If a peer is not responsible for a Resource-ID k, but is directly 4711 connected to a node with Node-ID k, then it routes the message to 4712 that node. Otherwise, it routes the request to the peer in the 4713 routing table that has the largest Node-ID that is in the interval 4714 between the peer and k. If no such node is found, it finds the 4715 smallest Node-Id that is greater than k and routes the message to 4716 that node. 4718 9.4. Redundancy 4720 When a peer receives a Store request for Resource-ID k, and it is 4721 responsible for Resource-ID k, it stores the data and returns a 4722 success response. It then sends a Store request to its successor in 4723 the neighbor table and to that peer's successor. Note that these 4724 Store requests are addressed to those specific peers, even though the 4725 Resource-ID they are being asked to store is outside the range that 4726 they are responsible for. The peers receiving these check they came 4727 from an appropriate predecessor in their neighbor table and that they 4728 are in a range that this predecessor is responsible for, and then 4729 they store the data. They do not themselves perform further Stores 4730 because they can determine that they are not responsible for the 4731 Resource-ID. 4733 Managing replicas as the overlay changes is described in 4734 Section 9.7.3. 4736 The sequential replicas used in this overlay algorithm protect 4737 against peer failure but not against malicious peers. Additional 4738 replication from the Usage is required to protect resources from such 4739 attacks, as discussed in Section 12.5.4. 4741 9.5. Joining 4743 The join process for a joining party (JP) with Node-ID n is as 4744 follows. 4746 1. JP MUST connect to its chosen bootstrap node. 4747 2. JP SHOULD send an Attach request to the admitting peer (AP) for 4748 Node-ID n. The "send_update" flag should be used to acquire the 4749 routing table for AP. 4750 3. JP SHOULD send Attach requests to initiate connections to each of 4751 the peers in the neighbor table as well as to the desired finger 4752 table entries. Note that this does not populate their routing 4753 tables, but only their connection tables, so JP will not get 4754 messages that it is expected to route to other nodes. 4755 4. JP MUST enter all the peers it has contacted into its routing 4756 table. 4757 5. JP MUST send a Join to AP. The AP sends the response to the 4758 Join. 4759 6. AP MUST do a series of Store requests to JP to store the data 4760 that JP will be responsible for. 4761 7. AP MUST send JP an Update explicitly labeling JP as its 4762 predecessor. At this point, JP is part of the ring and 4763 responsible for a section of the overlay. AP can now forget any 4764 data which is assigned to JP and not AP. 4766 8. The AP MUST send an Update to all of its neighbors with the new 4767 values of its neighbor set (including JP). 4768 9. The JP MUST send Updates to all the peers in its neighbor table. 4770 If JP sends an Attach to AP with send_update, it immediately knows 4771 most of its expected neighbors from AP's routing table update and can 4772 directly connect to them. This is the RECOMMENDED procedure. 4774 If for some reason JP does not get AP's routing table, it can still 4775 populate its neighbor table incrementally. It sends a Ping directed 4776 at Resource-ID n+1 (directly after its own Resource-ID). This allows 4777 it to discover its own successor. Call that node p0. It then sends 4778 a ping to p0+1 to discover its successor (p1). This process can be 4779 repeated to discover as many successors as desired. The values for 4780 the two peers before p will be found at a later stage when n receives 4781 an Update. An alternate procedure is to send Attaches to those nodes 4782 rather than pings, which forms the connections immediately but may be 4783 slower if the nodes need to collect ICE candidates, thus reducing 4784 parallelism. 4786 In order to set up its finger table entry for peer i, JP simply sends 4787 an Attach to peer (n+2^(128-i). This will be routed to a peer in 4788 approximately the right location around the ring. 4790 The joining peer MUST NOT send any Update message placing itself in 4791 the overlay until it has successfully completed an Attach with each 4792 peer that should be in its neighbor table. 4794 9.6. Routing Attaches 4796 When a peer needs to Attach to a new peer in its neighbor table, it 4797 MUST source-route the Attach request through the peer from which it 4798 learned the new peer's Node-ID. Source-routing these requests allows 4799 the overlay to recover from instability. 4801 All other Attach requests, such as those for new finger table 4802 entries, are routed conventionally through the overlay. 4804 9.7. Updates 4806 An Update for this DHT is defined as 4807 enum { reserved (0), 4808 peer_ready(1), neighbors(2), full(3), (255) } 4809 ChordUpdateType; 4811 struct { 4812 uint32 uptime; 4813 ChordUpdateType type; 4814 select(type){ 4815 case peer_ready: /* Empty */ 4816 ; 4818 case neighbors: 4819 NodeId predecessors<0..2^16-1>; 4820 NodeId successors<0..2^16-1>; 4822 case full: 4823 NodeId predecessors<0..2^16-1>; 4824 NodeId successors<0..2^16-1>; 4825 NodeId fingers<0..2^16-1>; 4826 }; 4827 } ChordUpdate; 4829 The "uptime" field contains the time this peer has been up in 4830 seconds. 4832 The "type" field contains the type of the update, which depends on 4833 the reason the update was sent. 4835 peer_ready: this peer is ready to receive messages. This message 4836 is used to indicate that a node which has Attached is a peer and 4837 can be routed through. It is also used as a connectivity check to 4838 non-neighbor peers. 4840 neighbors: this version is sent to members of the Chord neighbor 4841 table. 4843 full: this version is sent to peers which request an Update with a 4844 RouteQueryReq. 4846 If the message is of type "neighbors", then the contents of the 4847 message will be: 4849 predecessors 4850 The predecessor set of the Updating peer. 4852 successors 4853 The successor set of the Updating peer. 4855 If the message is of type "full", then the contents of the message 4856 will be: 4858 predecessors 4859 The predecessor set of the Updating peer. 4861 successors 4862 The successor set of the Updating peer. 4864 fingers 4865 The finger table of the Updating peer, in numerically ascending 4866 order. 4868 A peer MUST maintain an association (via Attach) to every member of 4869 its neighbor set. A peer MUST attempt to maintain at least three 4870 predecessors and three successors, even though this will not be 4871 possible if the ring is very small. It is RECOMMENDED that O(log(N)) 4872 predecessors and successors be maintained in the neighbor set. 4874 9.7.1. Handling Neighbor Failures 4876 Every time a connection to a peer in the neighbor table is lost (as 4877 determined by connectivity pings or the failure of some request), the 4878 peer MUST remove the entry from its neighbor table and replace it 4879 with the best match it has from the other peers in its routing table. 4880 If using reactive recovery, it then sends an immediate Update to all 4881 nodes in its Neighbor Table. The update will contain all the Node- 4882 IDs of the current entries of the table (after the failed one has 4883 been removed). Note that when replacing a successor the peer SHOULD 4884 delay the creation of new replicas for successor replacement hold- 4885 down time (30 seconds) after removing the failed entry from its 4886 neighbor table in order to allow a triggered update to inform it of a 4887 better match for its neighbor table. 4889 If the neighbor failure effects the peer's range of responsible IDs, 4890 then the Update MUST be sent to all nodes in its Connection Table. 4892 A peer MAY attempt to reestablish connectivity with a lost neighbor 4893 either by waiting additional time to see if connectivity returns or 4894 by actively routing a new Attach to the lost peer. Details for these 4895 procedures are beyond the scope of this document. In no event does 4896 an attempt to reestablish connectivity with a lost neighbor allow the 4897 peer to remain in the neighbor table. Such a peer is returned to the 4898 neighbor table once connectivity is reestablished. 4900 If connectivity is lost to all successor peers in the neighbor table, 4901 then this peer should behave as if it is joining the network and use 4902 Pings to find a peer and send it a Join. If connectivity is lost to 4903 all the peers in the finger table, this peer should assume that it 4904 has been disconnected from the rest of the network, and it should 4905 periodically try to join the DHT. 4907 9.7.2. Handling Finger Table Entry Failure 4909 If a finger table entry is found to have failed, all references to 4910 the failed peer are removed from the finger table and replaced with 4911 the closest preceding peer from the finger table or neighbor table. 4913 If using reactive recovery, the peer initiates a search for a new 4914 finger table entry as described below. 4916 9.7.3. Receiving Updates 4918 When a peer, N, receives an Update request, it examines the Node-IDs 4919 in the UpdateReq and at its neighbor table and decides if this 4920 UpdateReq would change its neighbor table. This is done by taking 4921 the set of peers currently in the neighbor table and comparing them 4922 to the peers in the update request. There are two major cases: 4924 o The UpdateReq contains peers that match N's neighbor table, so no 4925 change is needed to the neighbor set. 4926 o The UpdateReq contains peers N does not know about that should be 4927 in N's neighbor table, i.e. they are closer than entries in the 4928 neighbor table. 4930 In the first case, no change is needed. 4932 In the second case, N MUST attempt to Attach to the new peers and if 4933 it is successful it MUST adjust its neighbor set accordingly. Note 4934 that it can maintain the now inferior peers as neighbors, but it MUST 4935 remember the closer ones. 4937 After any Pings and Attaches are done, if the neighbor table changes 4938 and the peer is using reactive recovery, the peer sends an Update 4939 request to each member of its Connection Table. These Update 4940 requests are what end up filling in the predecessor/successor tables 4941 of peers that this peer is a neighbor to. A peer MUST NOT enter 4942 itself in its successor or predecessor table and instead should leave 4943 the entries empty. 4945 If peer N is responsible for a Resource-ID R, and N discovers that 4946 the replica set for R (the next two nodes in its successor set) has 4947 changed, it MUST send a Store for any data associated with R to any 4948 new node in the replica set. It SHOULD NOT delete data from peers 4949 which have left the replica set. 4951 When a peer N detects that it is no longer in the replica set for a 4952 resource R (i.e., there are three predecessors between N and R), it 4953 SHOULD delete all data associated with R from its local store. 4955 When a peer discovers that its range of responsible IDs have changed, 4956 it MUST send an Update to all entries in its connection table. 4958 9.7.4. Stabilization 4960 There are four components to stabilization: 4961 1. exchange Updates with all peers in its neighbor table to exchange 4962 state. 4963 2. search for better peers to place in its finger table. 4964 3. search to determine if the current finger table size is 4965 sufficiently large. 4966 4. search to determine if the overlay has partitioned and needs to 4967 recover. 4969 9.7.4.1. Updating neighbor table 4971 A peer MUST periodically send an Update request to every peer in its 4972 Connection Table. The purpose of this is to keep the predecessor and 4973 successor lists up to date and to detect failed peers. The default 4974 time is about every ten minutes, but the configuration server SHOULD 4975 set this in the configuration document using the "chord-update- 4976 interval" element (denominated in seconds.) A peer SHOULD randomly 4977 offset these Update requests so they do not occur all at once. 4979 9.7.4.2. Refreshing finger table 4981 A peer MUST periodically search for new peers to replace invalid 4982 entries in the finger table. A finger table entry i is valid if it 4983 is in the range [ n+2^( 128-i ) , n+2^( 128-(i-1) )-1 ]. Invalid 4984 entries occur in the finger table when a previous finger table entry 4985 has failed or when no peer has been found in that range. 4987 A peer SHOULD NOT send Ping requests looking for new finger table 4988 entries more often than the configuration element "chord-ping- 4989 interval", which defaults to 3600 seconds (one per hour). 4991 Two possible methods for searching for new peers for the finger table 4992 entries are presented: 4994 Alternative 1: A peer selects one entry in the finger table from 4995 among the invalid entries. It pings for a new peer for that finger 4996 table entry. The selection SHOULD be exponentially weighted to 4997 attempt to replace earlier (lower i) entries in the finger table. A 4998 simple way to implement this selection is to search through the 4999 finger table entries from i=0 and each time an invalid entry is 5000 encountered, send a Ping to replace that entry with probability 0.5. 5002 Alternative 2: A peer monitors the Update messages received from its 5003 connections to observe when an Update indicates a peer that would be 5004 used to replace in invalid finger table entry, i, and flags that 5005 entry in the finger table. Every "chord-ping-interval" seconds, the 5006 peer selects from among those flagged candidates using an 5007 exponentially weighted probability as above. 5009 When searching for a better entry, the peer SHOULD send the Ping to a 5010 Node-ID selected randomly from that range. Random selection is 5011 preferred over a search for strictly spaced entries to minimize the 5012 effect of churn on overlay routing [minimizing-churn-sigcomm06]. An 5013 implementation or subsequent specification MAY choose a method for 5014 selecting finger table entries other than choosing randomly within 5015 the range. Any such alternate methods SHOULD be employed only on 5016 finger table stabilization and not for the selection of initial 5017 finger table entries unless the alternative method is faster and 5018 imposes less overhead on the overlay. 5020 A peer MAY choose to keep connections to multiple peers that can act 5021 for a given finger table entry. 5023 9.7.4.3. Adjusting finger table size 5025 If the finger table has less than 16 entries, the node SHOULD attempt 5026 to discover more fingers to grow the size of the table to 16. The 5027 value 16 was chosen to ensure high odds of a node maintaining 5028 connectivity to the overlay even with strange network partitions. 5030 For many overlays, 16 finger table entries will be enough, but as an 5031 overlay grows very large, more than 16 entries may be required in the 5032 finger table for efficient routing. An implementation SHOULD be 5033 capable of increasing the number of entries in the finger table to 5034 128 entries. 5036 Note to implementers: Although log(N) entries are all that are 5037 required for optimal performance, careful implementation of 5038 stabilization will result in no additional traffic being generated 5039 when maintaining a finger table larger than log(N) entries. 5040 Implementers are encouraged to make use of RouteQuery and algorithms 5041 for determining where new finger table entries may be found. 5042 Complete details of possible implementations are outside the scope of 5043 this specification. 5045 A simple approach to sizing the finger table is to ensure the finger 5046 table is large enough to contain at least the final successor in the 5047 peer's neighbor table. 5049 9.7.4.4. Detecting partitioning 5051 To detect that a partitioning has occurred and to heal the overlay, a 5052 peer P MUST periodically repeat the discovery process used in the 5053 initial join for the overlay to locate an appropriate bootstrap node, 5054 B. P should then send a Ping for its own Node-ID routed through B. If 5055 a response is received from a peer S', which is not P's successor, 5056 then the overlay is partitioned and P should send an Attach to S' 5057 routed through B, followed by an Update sent to S'. (Note that S' 5058 may not be in P's neighbor table once the overlay is healed, but the 5059 connection will allow S' to discover appropriate neighbor entries for 5060 itself via its own stabilization.) 5062 Future specifications may describe alternative mechanisms for 5063 determining when to repeat the discovery process. 5065 9.8. Route query 5067 For this topology plugin, the RouteQueryReq contains no additional 5068 information. The RouteQueryAns contains the single node ID of the 5069 next peer to which the responding peer would have routed the request 5070 message in recursive routing: 5072 struct { 5073 NodeId next_peer; 5074 } ChordRouteQueryAns; 5076 The contents of this structure are as follows: 5078 next_peer 5079 The peer to which the responding peer would route the message in 5080 order to deliver it to the destination listed in the request. 5082 If the requester has set the send_update flag, the responder SHOULD 5083 initiate an Update immediately after sending the RouteQueryAns. 5085 9.9. Leaving 5087 To support extensions, such as [I-D.ietf-p2psip-self-tuning], Peers 5088 SHOULD send a Leave request to all members of their neighbor table 5089 prior to exiting the Overlay Instance. The overlay_specific_data 5090 field MUST contain the ChordLeaveData structure defined below: 5092 enum { reserved (0), 5093 from_succ(1), from_pred(2), (255) } 5094 ChordLeaveType; 5096 struct { 5097 ChordLeaveType type; 5099 select(type) { 5100 case from_succ: 5101 NodeId successors<0..2^16-1>; 5102 case from_pred: 5103 NodeId predecessors<0..2^16-1>; 5104 }; 5105 } ChordLeaveData; 5107 The 'type' field indicates whether the Leave request was sent by a 5108 predecessor or a successor of the recipient: 5110 from_succ 5111 The Leave request was sent by a successor. 5113 from_pred 5114 The Leave request was sent by a predecessor. 5116 If the type of the request is 'from_succ', the contents will be: 5118 successors 5119 The sender's successor list. 5121 If the type of the request is 'from_pred', the contents will be: 5123 predecessors 5124 The sender's predecessor list. 5126 Any peer which receives a Leave for a peer n in its neighbor set 5127 follows procedures as if it had detected a peer failure as described 5128 in Section 9.7.1. 5130 10. Enrollment and Bootstrap 5132 The section defines the format of the configuration data as well the 5133 process to join a new overlay. 5135 10.1. Overlay Configuration 5137 This specification defines a new content type "application/ 5138 p2p-overlay+xml" for an MIME entity that contains overlay 5139 information. An example document is shown below. 5141 5142 5145 5147 CHORD-RELOAD 5148 16 5149 5150 MIIDJDCCAo2gAwIBAgIBADANBgkqhkiG9w0BAQUFADBwMQswCQYDVQQGEwJVUzET 5151 MBEGA1UECBMKQ2FsaWZvcm5pYTERMA8GA1UEBxMIU2FuIEpvc2UxDjAMBgNVBAoT 5152 BXNpcGl0MSkwJwYDVQQLEyBTaXBpdCBUZXN0IENlcnRpZmljYXRlIEF1dGhvcml0 5153 eTAeFw0wMzA3MTgxMjIxNTJaFw0xMzA3MTUxMjIxNTJaMHAxCzAJBgNVBAYTAlVT 5154 MRMwEQYDVQQIEwpDYWxpZm9ybmlhMREwDwYDVQQHEwhTYW4gSm9zZTEOMAwGA1UE 5155 ChMFc2lwaXQxKTAnBgNVBAsTIFNpcGl0IFRlc3QgQ2VydGlmaWNhdGUgQXV0aG9y 5156 aXR5MIGfMA0GCSqGSIb3DQEBAQUAA4GNADCBiQKBgQDDIh6DkcUDLDyK9BEUxkud 5157 +nJ4xrCVGKfgjHm6XaSuHiEtnfELHM+9WymzkBNzZpJu30yzsxwfKoIKugdNUrD4 5158 N3viCicwcN35LgP/KnbN34cavXHr4ZlqxH+OdKB3hQTpQa38A7YXdaoz6goW2ft5 5159 Mi74z03GNKP/G9BoKOGd5QIDAQABo4HNMIHKMB0GA1UdDgQWBBRrRhcU6pR2JYBU 5160 bhNU2qHjVBShtjCBmgYDVR0jBIGSMIGPgBRrRhcU6pR2JYBUbhNU2qHjVBShtqF0 5161 pHIwcDELMAkGA1UEBhMCVVMxEzARBgNVBAgTCkNhbGlmb3JuaWExETAPBgNVBAcT 5162 CFNhbiBKb3NlMQ4wDAYDVQQKEwVzaXBpdDEpMCcGA1UECxMgU2lwaXQgVGVzdCBD 5163 ZXJ0aWZpY2F0ZSBBdXRob3JpdHmCAQAwDAYDVR0TBAUwAwEB/zANBgkqhkiG9w0B 5164 AQUFAAOBgQCWbRvv1ZGTRXxbH8/EqkdSCzSoUPrs+rQqR0xdQac9wNY/nlZbkR3O 5165 qAezG6Sfmklvf+DOg5RxQq/+Y6I03LRepc7KeVDpaplMFGnpfKsibETMipwzayNQ 5166 QgUf4cKBiF+65Ue7hZuDJa2EMv8qW4twEhGDYclpFU9YozyS1OhvUg== 5167 5168 YmFkIGNlcnQK 5169 https://example.org 5170 https://example.net 5171 false 5173 5174 5175 5176 20 5177 5178 5179 false 5180 false 5181 5182 400 5183 30 5184 true 5185 password 5186 4000 5187 30 5188 TLS 5189 47112162e84c69ba 5190 47112162e84c69ba 5191 6eba45d31a900c06 5192 6ebc45d31a900c06 5193 6ebc45d31a900ca6 5195 foo 5197 5198 urn:ietf:params:xml:ns:p2p:config-ext1 5199 5201 5202 5203 5204 SINGLE 5205 USER-MATCH 5206 1 5207 100 5208 5209 5210 VGhpcyBpcyBub3QgcmlnaHQhCg== 5211 5212 5213 5214 5215 ARRAY 5216 NODE-MULTIPLE 5217 3 5218 22 5219 4 5220 1 5221 5222 5223 5224 VGhpcyBpcyBub3QgcmlnaHQhCg== 5225 5227 5228 5229 5230 VGhpcyBpcyBub3QgcmlnaHQhCg== 5232 5233 5234 VGhpcyBpcyBub3QgcmlnaHQhCg== 5236 5238 The file MUST be a well formed XML document and it SHOULD contain an 5239 encoding declaration in the XML declaration. The file MUST use the 5240 UTF-8 character encoding. The namespace for the elements defined in 5241 this specification is urn:ietf:params:xml:ns:p2p:config-base and 5242 urn:ietf:params:xml:ns:p2p:config-chord". 5244 The file can contain multiple "configuration" elements where each one 5245 contains the configuration information for a different overlay. Each 5246 configuration element may be followed by signature elements that 5247 provides a signature over the preceding configuration element. Each 5248 configuration element has the following attributes: 5250 instance-name: name of the overlay 5251 expiration: time in the future at which this overlay configuration 5252 is no longer valid. The node SHOULD retrieve a new copy of the 5253 configuration at a randomly selected time that is before the 5254 expiration time. Note that if the certificates expire before a 5255 new configuration is retried, the node will not be able to 5256 validate the configuration file. 5257 sequence: a monotonically increasing sequence number between 1 and 5258 2^16-2 5260 Inside each overlay element, the following elements can occur: 5262 topology-plugin This element defines the overlay algorithm being 5263 used. If missing the default is "CHORD-RELOAD". 5264 node-id-length This element contains the length of a NodeId 5265 (NodeIdLength) in bytes. This value MUST be between 16 (128 bits) 5266 and 20 (160 bits). If this element is not present, the default of 5267 16 is used. 5268 root-cert This element contains a base-64 encoded X.509v3 5269 certificate that is a root trust anchor used to sign all 5270 certificates in this overlay. There can be more than one root- 5271 cert element. 5273 enrollment-server This element contains the URL at which the 5274 enrollment server can be reached in a "url" element. This URL 5275 MUST be of type "https:". More than one enrollment-server element 5276 may be present. 5277 self-signed-permitted This element indicates whether self-signed 5278 certificates are permitted. If it is set to "true", then self- 5279 signed certificates are allowed, in which case the enrollment- 5280 server and root-cert elements may be absent. Otherwise, it SHOULD 5281 be absent, but MAY be set to "false". This element also contains 5282 an attribute "digest" which indicates the digest to be used to 5283 compute the Node-ID. Valid values for this parameter are "sha1" 5284 and "sha256" representing SHA-1 [RFC3174] and SHA-256 [RFC6234] 5285 respectively. Implementations MUST support both of these 5286 algorithms. 5287 bootstrap-node This element represents the address of one of the 5288 bootstrap nodes. It has an attribute called "address" that 5289 represents the IP address (either IPv4 or IPv6, since they can be 5290 distinguished) and an optional attribute called "port" that 5291 represents the port and defaults to 6084. The IP address is in 5292 typical hexadecimal form using standard period and colon 5293 separators as specified in [RFC5952]. More than one bootstrap- 5294 peer element may be present. 5295 turn-density This element is a positive integer that represents the 5296 approximate reciprocal of density of nodes that can act as TURN 5297 servers. For example, if 5% of the nodes can act as TURN servers, 5298 this would be set to 20. If it is not present, the default value 5299 is 1. If there are no TURN servers in the overlay, it is set to 5300 zero. 5301 multicast-bootstrap This element represents the address of a 5302 multicast, broadcast, or anycast address and port that may be used 5303 for bootstrap. Nodes SHOULD listen on the address. It has an 5304 attributed called "address" that represents the IP address and an 5305 optional attribute called "port" that represents the port and 5306 defaults to 6084. More than one "multicast-bootstrap" element may 5307 be present. 5308 clients-permitted This element represents whether clients are 5309 permitted or whether all nodes must be peers. If it is set to 5310 "true" or absent, this indicates that clients are permitted. If 5311 it is set to "false" then nodes are not allowed to remain clients 5312 after the initial join. There is currently no way for the overlay 5313 to enforce this. 5314 no-ice This element represents whether nodes are required to use 5315 the "No-ICE" Overlay Link protocols in this overlay. If it is 5316 absent, it is treated as if it were set to "false". 5318 chord-update-interval The update frequency for the Chord-reload 5319 topology plugin (see Section 9). 5320 chord-ping-interval The ping frequency for the Chord-reload 5321 topology plugin (see Section 9). 5322 chord-reactive Whether reactive recovery should be used for this 5323 overlay. Set to "true" or "false". Default if missing is "true". 5324 (see Section 9). 5325 shared-secret If shared secret mode is used, this contains the 5326 shared secret. 5327 max-message-size Maximum size in bytes of any message in the 5328 overlay. If this value is not present, the default is 5000. 5329 initial-ttl Initial default TTL (time to live, see Section 5.3.2) 5330 for messages. If this value is not present, the default is 100. 5331 overlay-link-protocol Indicates a permissible overlay link protocol 5332 (see Section 5.6.1 for requirements for such protocols). An 5333 arbitrary number of these elements may appear. If none appear, 5334 then this implies the default value, "TLS", which refers to the 5335 use of TLS and DTLS. If one or more elements appear, then no 5336 default value applies. 5337 kind-signer This contains a single Node-ID in hexadecimal and 5338 indicates that the certificate with this Node-ID is allowed to 5339 sign kinds. Identifying kind-signer by Node-ID instead of 5340 certificate allows the use of short lived certificates without 5341 constantly having to provide an updated configuration file. 5342 configuration-signer This contains a single Node-ID in hexadecimal 5343 and indicates that the certificate with this Node-ID is allowed to 5344 sign configurations for this instance-name. Identifying the 5345 signer by Node-ID instead of certificate allows the use of short 5346 lived certificates without constantly having to provide an updated 5347 configuration file. 5348 bad-node This contains a single Node-ID in hexadecimal and 5349 indicates that the certificate with this Node-ID MUST NOT be 5350 considered valid. This allows certificate revocation. An 5351 arbitrary number of these elements can be provided. Note that 5352 because certificates may expire, bad-node entries need only be 5353 present for the lifetime of the certificate. Technically 5354 speaking, bad node-ids may be reused once their certificates have 5355 expired, the requirement for node-ids to be pseudo randomly 5356 generated gives this event a vanishing probability. 5357 mandatory-extension This element contains the name of an XML 5358 namespace that a node joining the overlay MUST support. The 5359 presence of a mandatory-extension element does not require the 5360 extension to be used in the current configuration file, but can 5361 indicate that it may be used in the future. Note that the 5362 namespace is case-sensitive, as specified in [w3c-xml-namespaces] 5363 Section 2.3. More than one mandatory-extension element may be 5364 present. 5366 Inside each overlay element, the required-kinds elements can also 5367 occur. This element indicates the kinds that members must support 5368 and contains multiple kind-block elements that each define a single 5369 kind that MUST be supported by nodes in the overlay. Each kind-block 5370 consists of a single kind element and a kind-signature. The kind 5371 element defines the kind. The kind-signature is the signature 5372 computed over the kind element. 5374 Each kind has either an id attribute or a name attribute. The name 5375 attribute is a string representing the kind (the name registered to 5376 IANA) while the id is an integer kind-id allocated out of private 5377 space. 5379 In addition, the kind element contains the following elements: 5380 max-count: the maximum number of values which members of the overlay 5381 must support. 5382 data-model: the data model to be used. 5383 max-size: the maximum size of individual values. 5384 access-control: the access control model to be used. 5385 max-node-multiple: This is optional and only used when the access 5386 control is NODE-MULTIPLE. This indicates the maximum value for 5387 the i counter. This is an integer greater than 0. 5389 All of the non optional values MUST be provided. If the kind is 5390 registered with IANA, the data-model and access-control elements MUST 5391 match those in the kind registration, and clients MUST ignore them in 5392 favor of the IANA versions. Multiple required-kinds elements MAY be 5393 present. 5395 The kind-block element also MUST contain a "kind-signature" element. 5396 This signature is computed across the kind from the beginning of the 5397 first < of the kind to the end of the last > of the kind in the same 5398 way as the signature element described later in this section. 5400 The configuration file is a binary file and cannot be changed - 5401 including whitespace changes - or the signature will break. The 5402 signature is computed by taking each configuration element and 5403 starting from, and including, the first < at the start of 5404 up to and including the > in and 5405 treating this as a binary blob that is signed using the standard 5406 SecurityBlock defined in Section 5.3.4. The SecurityBlock is base 64 5407 encoded using the base64 alphabet from RFC[RFC4648] and put in the 5408 signature element following the configuration object in the 5409 configuration file. 5411 When a node receives a new configuration file, it MUST change its 5412 configuration to meet the new requirements. This may require the 5413 node to exit the DHT and re-join. If a node is not capable of 5414 supporting the new requirements, it MUST exit the overlay. If some 5415 information about a particular kind changes from what the node 5416 previously knew about the kind (for example the max size), the new 5417 information in the configuration files overrides any previously 5418 learned information. If any kind data was signed by a node that is 5419 no longer allowed to sign kinds, that kind MUST be discarded along 5420 with any stored information of that kind. Note that forcing an 5421 avalanche restart of the overlay with a configuration change that 5422 requires re-joining the overlay may result in serious performance 5423 problems, including total collapse of the network if configuration 5424 parameters are not properly considered. Such an event may be 5425 necessary in case of a compromised CA or similar problem, but for 5426 large overlays should be avoided in almost all circumstances. 5428 10.1.1. Relax NG Grammar 5430 The grammar for the configuration data is: 5432 namespace chord = "urn:ietf:params:xml:ns:p2p:config-chord" 5433 namespace local = "" 5434 default namespace p2pcf = "urn:ietf:params:xml:ns:p2p:config-base" 5435 namespace rng = "http://relaxng.org/ns/structure/1.0" 5437 anything = 5438 (element * { anything } 5439 | attribute * { text } 5440 | text)* 5442 foreign-elements = element * - (p2pcf:* | local:* | chord:*) 5443 { anything }* 5444 foreign-attributes = attribute * - (p2pcf:*|local:*|chord:*) 5445 { text }* 5446 foreign-nodes = (foreign-attributes | foreign-elements)* 5448 start = element p2pcf:overlay { 5449 overlay-element 5450 } 5452 overlay-element &= element configuration { 5453 attribute instance-name { xsd:string }, 5454 attribute expiration { xsd:dateTime }?, 5455 attribute sequence { xsd:long }?, 5456 foreign-attributes*, 5457 parameter 5458 }+ 5459 overlay-element &= element signature { 5460 attribute algorithm { signature-algorithm-type }?, 5461 xsd:base64Binary 5463 }* 5465 signature-algorithm-type |= "rsa-sha1" 5466 signature-algorithm-type |= xsd:string # signature alg extensions 5468 parameter &= element topology-plugin { topology-plugin-type }? 5469 topology-plugin-type |= xsd:string # topo plugin extensions 5470 parameter &= element max-message-size { xsd:unsignedInt }? 5471 parameter &= element initial-ttl { xsd:int }? 5472 parameter &= element root-cert { xsd:base64Binary }* 5473 parameter &= element required-kinds { kind-block* }? 5474 parameter &= element enrollment-server { xsd:anyURI }* 5475 parameter &= element kind-signer { xsd:string }* 5476 parameter &= element configuration-signer { xsd:string }* 5477 parameter &= element bad-node { xsd:string }* 5478 parameter &= element no-ice { xsd:boolean }? 5479 parameter &= element shared-secret { xsd:string }? 5480 parameter &= element overlay-link-protocol { xsd:string }* 5481 parameter &= element clients-permitted { xsd:boolean }? 5482 parameter &= element turn-density { xsd:unsignedByte }? 5483 parameter &= element node-id-length { xsd:int }? 5484 parameter &= element mandatory-extension { xsd:string }* 5485 parameter &= foreign-elements* 5487 parameter &= 5488 element self-signed-permitted { 5489 attribute digest { self-signed-digest-type }, 5490 xsd:boolean 5491 }? 5492 self-signed-digest-type |= "sha1" 5493 self-signed-digest-type |= xsd:string # signature digest extensions 5495 parameter &= element bootstrap-node { 5496 attribute address { xsd:string }, 5497 attribute port { xsd:int }? 5498 }* 5500 parameter &= element multicast-bootstrap { 5501 attribute address { xsd:string }, 5502 attribute port { xsd:int }? 5503 }* 5505 kind-block = element kind-block { 5506 element kind { 5507 ( attribute name { kind-names } 5508 | attribute id { xsd:unsignedInt } ), 5509 kind-parameter 5510 } & 5511 element kind-signature { 5512 attribute algorithm { signature-algorithm-type }?, 5513 xsd:base64Binary 5514 }? 5515 } 5517 kind-parameter &= element max-count { xsd:int } 5518 kind-parameter &= element max-size { xsd:int } 5519 kind-parameter &= element max-node-multiple { xsd:int }? 5521 kind-parameter &= element data-model { data-model-type } 5522 data-model-type |= "SINGLE" 5523 data-model-type |= "ARRAY" 5524 data-model-type |= "DICTIONARY" 5525 data-model-type |= xsd:string # data model extensions 5527 kind-parameter &= element access-control { access-control-type } 5528 access-control-type |= "USER-MATCH" 5529 access-control-type |= "NODE-MATCH" 5530 access-control-type |= "USER-NODE-MATCH" 5531 access-control-type |= "NODE-MULTIPLE" 5532 access-control-type |= xsd:string # access control extensions 5534 kind-parameter &= foreign-elements* 5536 kind-names |= "TURN-SERVICE" 5537 kind-names |= "CERTIFICATE_BY_NODE" 5538 kind-names |= "CERTIFICATE_BY_USER" 5539 kind-names |= xsd:string # kind extensions 5541 # Chord specific parameters 5542 topology-plugin-type |= "CHORD-RELOAD" 5543 parameter &= element chord:chord-ping-interval { xsd:int }? 5544 parameter &= element chord:chord-update-interval { xsd:int }? 5545 parameter &= element chord:chord-reactive { xsd:boolean }? 5547 10.2. Discovery Through Configuration Server 5549 When a node first enrolls in a new overlay, it starts with a 5550 discovery process to find a configuration server. 5552 The node MAY start by determines the overlay name. This value is 5553 provided by the user or some other out of band provisioning 5554 mechanism. The out of band mechanisms MAY also provide an optional 5555 URL for the configuration server. If a URL for the configuration 5556 server is not provided, the node MUST do a DNS SRV query using a 5557 Service name of "p2psip-enroll" and a protocol of TCP to find a 5558 configuration server and form the URL by appending a path of "/.well- 5559 known/p2psip-enroll" to the overlay name. This uses the "well known 5560 URI" framework defined in [RFC5785]. For example, if the overlay 5561 name was example.com, the URL would be 5562 "https://example.com//.well-known/p2psip-enroll". 5564 Once an address and URL for the configuration server is determined, 5565 the peer forms an HTTPS connection to that IP address. The 5566 certificate MUST match the overlay name as described in [RFC2818]. 5567 Then the node MUST fetch a new copy of the configuration file. To do 5568 this, the peer performs a GET to the URL. The result of the HTTP GET 5569 is an XML configuration file described above, which replaces any 5570 previously learned configuration file for this overlay. 5572 For overlays that do not use a configuration server, nodes obtain the 5573 configuration information needed to join the overlay through some out 5574 of band approach such an XML configuration file sent over email. 5576 10.3. Credentials 5578 If the configuration document contains a enrollment-server element, 5579 credentials are required to join the Overlay Instance. A peer which 5580 does not yet have credentials MUST contact the enrollment server to 5581 acquire them. 5583 RELOAD defines its own trivial certificate request protocol. We 5584 would have liked to have used an existing protocol but were concerned 5585 about the implementation burden of even the simplest of those 5586 protocols, such as [RFC5272] and [RFC5273]. Our objective was to 5587 have a protocol which could be easily implemented in a Web server 5588 which the operator did not control (e.g., in a hosted service) and 5589 was compatible with the existing certificate handling tooling as used 5590 with the Web certificate infrastructure. This means accepting bare 5591 PKCS#10 requests and returning a single bare X.509 certificate. 5592 Although the MIME types for these objects are defined, none of the 5593 existing protocols support exactly this model. 5595 The certificate request protocol is performed over HTTPS. The 5596 request is an HTTP POST with the following properties: 5598 o If authentication is required, there is an URL parameter of 5599 "password" and "username" containing the user's name and password 5600 in the clear (hence the need for HTTPS) 5601 o If more than one Node-ID is required, there is an URL parameter of 5602 "nodeids" containing the number of Node-IDs required. 5603 o The body is of content type "application/pkcs10", as defined in 5604 [RFC2311]. 5606 o The Accept header contains the type "application/pkix-cert", 5607 indicating the type that is expected in the response. 5609 The enrollment server MUST authenticate the request using the 5610 provided user name and password. If the authentication succeeds and 5611 the requested user name is acceptable, the server generates and 5612 returns a certificate. The SubjectAltName field in the certificate 5613 contains the following values: 5615 o One or more Node-IDs which MUST be cryptographically random 5616 [RFC4086]. Each MUST be chosen by the enrollment server in such a 5617 way that they are unpredictable to the requesting user. E.g., the 5618 user MUST NOT be informed of potential (random) Node-IDs prior to 5619 authenticating. Each is placed in the subjectAltName using the 5620 uniformResourceIdentifier type and MUST contain RELOAD URIs as 5621 described in Section 13.15 and MUST contain a Destination list 5622 with a single entry of type "node_id". 5623 o A single name this user is allowed to use in the overlay, using 5624 type rfc822Name. 5626 The certificate is returned as type "application/pkix-cert" as 5627 defined in [RFC2585], with an HTTP status code of 200 OK. 5628 Certificate processing errors should be treated as HTTP errors and 5629 have appropriate HTTP status codes. 5631 The client MUST check that the certificate returned was signed by one 5632 of the certificates received in the "root-cert" list of the overlay 5633 configuration data. The node then reads the certificate to find the 5634 Node-IDs it can use. 5636 10.3.1. Self-Generated Credentials 5638 If the "self-signed-permitted" element is present in the 5639 configuration and set to "true", then a node MUST generate its own 5640 self-signed certificate to join the overlay. The self-signed 5641 certificate MAY contain any user name of the users choice. 5643 The Node-ID MUST be computed by applying the digest specified in the 5644 self-signed-permitted element to the DER representation of the user's 5645 public key (more specifically the subjectPublicKeyInfo) and taking 5646 the high order bits. When accepting a self-signed certificate, nodes 5647 MUST check that the Node-ID and public keys match. This prevents 5648 Node-ID theft. 5650 Once the node has constructed a self-signed certificate, it MAY join 5651 the overlay. Before storing its certificate in the overlay 5652 (Section 7) it SHOULD look to see if the user name is already taken 5653 and if so choose another user name. Note that this only provides 5654 protection against accidental name collisions. Name theft is still 5655 possible. If protection against name theft is desired, then the 5656 enrollment service must be used. 5658 10.4. Searching for a Bootstrap Node 5660 If no cached bootstrap nodes are available and the configuration file 5661 has an multicast-bootstrap element, then the node SHOULD send a Ping 5662 request over UDP to the address and port found to each multicast- 5663 bootstrap element found in the configuration document. This MAY be a 5664 multicast, broadcast, or anycast address. The Ping should use the 5665 wildcard Node-ID as the destination Node-ID. 5667 The responder node that receives the Ping request SHOULD check that 5668 the overlay name is correct and that the requester peer sending the 5669 request has appropriate credentials for the overlay before responding 5670 to the Ping request even if the response is only an error. 5672 10.5. Contacting a Bootstrap Node 5674 In order to join the overlay, the joining node MUST contact a node in 5675 the overlay. Typically this means contacting the bootstrap nodes, 5676 since they are reachable by the local peer or have public IP 5677 addresses. If the joining node has cached a list of peers it has 5678 previously been connected with in this overlay, as an optimization it 5679 MAY attempt to use one or more of them as bootstrap nodes before 5680 falling back to the bootstrap nodes listed in the configuration file. 5682 When contacting a bootstrap node, the joining node first forms the 5683 DTLS or TLS connection to the bootstrap node and then sends an Attach 5684 request over this connection with the destination Node-ID set to the 5685 joining node's Node-ID. 5687 When the requester node finally does receive a response from some 5688 responding node, it can note the Node-ID in the response and use this 5689 Node-ID to start sending requests to join the Overlay Instance as 5690 described in Section 5.4. 5692 After a node has successfully joined the overlay network, it will 5693 have direct connections to several peers. Some MAY be added to the 5694 cached bootstrap nodes list and used in future boots. Peers that are 5695 not directly connected MUST NOT be cached. The suggested number of 5696 peers to cache is 10. Algorithms for determining which peers to 5697 cache are beyond the scope of this specification. 5699 11. Message Flow Example 5701 The following abbreviation are used in the message flow diagrams: JP 5702 = joining peer, AP = admitting peer, NP = next peer after the AP, NNP 5703 = next next peer which is the peer after NP, PP = previous peer 5704 before the AP, PPP = previous previous peer which is the peer before 5705 the PP, BP = bootstrap peer. 5707 In the following example, we assume that JP has formed a connection 5708 to one of the bootstrap nodes. JP then sends an Attach through that 5709 peer to a resource ID of itself (JP). It gets routed to the 5710 admitting peer (AP) because JP is not yet part of the overlay. When 5711 AP responds, JP and AP use ICE to set up a connection and then set up 5712 TLS. Once AP has connected to JP, AP sends to JP an Update to 5713 populate its Routing Table. The following example shows the Update 5714 happening after the TLS connection is formed but it could also happen 5715 before in which case the Update would often be routed through other 5716 nodes. 5718 JP PPP PP AP NP NNP BP 5719 | | | | | | | 5720 | | | | | | | 5721 | | | | | | | 5722 |Attach Dest=JP | | | | | 5723 |---------------------------------------------------------->| 5724 | | | | | | | 5725 | | | | | | | 5726 | | |Attach Dest=JP | | | 5727 | | |<--------------------------------------| 5728 | | | | | | | 5729 | | | | | | | 5730 | | |Attach Dest=JP | | | 5731 | | |-------->| | | | 5732 | | | | | | | 5733 | | | | | | | 5734 | | |AttachAns | | | 5735 | | |<--------| | | | 5736 | | | | | | | 5737 | | | | | | | 5738 | | |AttachAns | | | 5739 | | |-------------------------------------->| 5740 | | | | | | | 5741 | | | | | | | 5742 |AttachAns | | | | | 5743 |<----------------------------------------------------------| 5744 | | | | | | | 5745 | | | | | | | 5746 |TLS | | | | | | 5747 |.............................| | | | 5748 | | | | | | | 5749 | | | | | | | 5750 | | | | | | | 5751 |Update | | | | | | 5752 |<----------------------------| | | | 5753 | | | | | | | 5754 | | | | | | | 5755 |UpdateAns| | | | | | 5756 |---------------------------->| | | | 5757 | | | | | | | 5758 | | | | | | | 5759 | | | | | | | 5761 The JP then forms connections to the appropriate neighbors, such as 5762 NP, by sending an Attach which gets routed via other nodes. When NP 5763 responds, JP and NP use ICE and TLS to set up a connection. 5765 JP PPP PP AP NP NNP BP 5766 | | | | | | | 5767 | | | | | | | 5768 | | | | | | | 5769 |Attach NP | | | | | 5770 |---------------------------->| | | | 5771 | | | | | | | 5772 | | | | | | | 5773 | | | |Attach NP| | | 5774 | | | |-------->| | | 5775 | | | | | | | 5776 | | | | | | | 5777 | | | |AttachAns| | | 5778 | | | |<--------| | | 5779 | | | | | | | 5780 | | | | | | | 5781 |AttachAns | | | | | 5782 |<----------------------------| | | | 5783 | | | | | | | 5784 | | | | | | | 5785 |Attach | | | | | | 5786 |-------------------------------------->| | | 5787 | | | | | | | 5788 | | | | | | | 5789 |TLS | | | | | | 5790 |.......................................| | | 5791 | | | | | | | 5792 | | | | | | | 5793 | | | | | | | 5794 | | | | | | | 5796 JP also needs to populate its finger table (for the Chord based DHT). 5797 It issues an Attach to a variety of locations around the overlay. 5798 The diagram below shows it sending an Attach halfway around the Chord 5799 ring to the JP + 2^127. 5801 JP NP XX TP 5802 | | | | 5803 | | | | 5804 | | | | 5805 |Attach JP+2<<126 | | 5806 |-------->| | | 5807 | | | | 5808 | | | | 5809 | |Attach JP+2<<126 | 5810 | |-------->| | 5811 | | | | 5812 | | | | 5813 | | |Attach JP+2<<126 5814 | | |-------->| 5815 | | | | 5816 | | | | 5817 | | |AttachAns| 5818 | | |<--------| 5819 | | | | 5820 | | | | 5821 | |AttachAns| | 5822 | |<--------| | 5823 | | | | 5824 | | | | 5825 |AttachAns| | | 5826 |<--------| | | 5827 | | | | 5828 | | | | 5829 |TLS | | | 5830 |.............................| 5831 | | | | 5832 | | | | 5833 | | | | 5834 | | | | 5836 Once JP has a reasonable set of connections, it is ready to take its 5837 place in the DHT. It does this by sending a Join to AP. AP does a 5838 series of Store requests to JP to store the data that JP will be 5839 responsible for. AP then sends JP an Update explicitly labeling JP 5840 as its predecessor. At this point, JP is part of the ring and 5841 responsible for a section of the overlay. AP can now forget any data 5842 which is assigned to JP and not AP. 5844 JP PPP PP AP NP NNP BP 5845 | | | | | | | 5846 | | | | | | | 5847 | | | | | | | 5848 |JoinReq | | | | | | 5849 |---------------------------->| | | | 5850 | | | | | | | 5851 | | | | | | | 5852 |JoinAns | | | | | | 5853 |<----------------------------| | | | 5854 | | | | | | | 5855 | | | | | | | 5856 |StoreReq Data A | | | | | 5857 |<----------------------------| | | | 5858 | | | | | | | 5859 | | | | | | | 5860 |StoreAns | | | | | | 5861 |---------------------------->| | | | 5862 | | | | | | | 5863 | | | | | | | 5864 |StoreReq Data B | | | | | 5865 |<----------------------------| | | | 5866 | | | | | | | 5867 | | | | | | | 5868 |StoreAns | | | | | | 5869 |---------------------------->| | | | 5870 | | | | | | | 5871 | | | | | | | 5872 |UpdateReq| | | | | | 5873 |<----------------------------| | | | 5874 | | | | | | | 5875 | | | | | | | 5876 |UpdateAns| | | | | | 5877 |---------------------------->| | | | 5878 | | | | | | | 5879 | | | | | | | 5880 | | | | | | | 5881 | | | | | | | 5883 In Chord, JP's neighbor table needs to contain its own predecessors. 5884 It couldn't connect to them previously because it did not yet know 5885 their addresses. However, now that it has received an Update from 5886 AP, it has AP's predecessors, which are also its own, so it sends 5887 Attaches to them. Below it is shown connecting to AP's closest 5888 predecessor, PP. 5890 JP PPP PP AP NP NNP BP 5891 | | | | | | | 5892 | | | | | | | 5893 | | | | | | | 5894 |Attach Dest=PP | | | | | 5895 |---------------------------->| | | | 5896 | | | | | | | 5897 | | | | | | | 5898 | | |Attach Dest=PP | | | 5899 | | |<--------| | | | 5900 | | | | | | | 5901 | | | | | | | 5902 | | |AttachAns| | | | 5903 | | |-------->| | | | 5904 | | | | | | | 5905 | | | | | | | 5906 |AttachAns| | | | | | 5907 |<----------------------------| | | | 5908 | | | | | | | 5909 | | | | | | | 5910 |TLS | | | | | | 5911 |...................| | | | | 5912 | | | | | | | 5913 | | | | | | | 5914 |UpdateReq| | | | | | 5915 |------------------>| | | | | 5916 | | | | | | | 5917 | | | | | | | 5918 |UpdateAns| | | | | | 5919 |<------------------| | | | | 5920 | | | | | | | 5921 | | | | | | | 5922 |UpdateReq| | | | | | 5923 |---------------------------->| | | | 5924 | | | | | | | 5925 | | | | | | | 5926 |UpdateAns| | | | | | 5927 |<----------------------------| | | | 5928 | | | | | | | 5929 | | | | | | | 5930 |UpdateReq| | | | | | 5931 |-------------------------------------->| | | 5932 | | | | | | | 5933 | | | | | | | 5934 |UpdateAns| | | | | | 5935 |<--------------------------------------| | | 5936 | | | | | | | 5937 | | | | | | | 5939 Finally, now that JP has a copy of all the data and is ready to route 5940 messages and receive requests, it sends Updates to everyone in its 5941 Routing Table to tell them it is ready to go. Below, it is shown 5942 sending such an update to TP. 5944 JP NP XX TP 5945 | | | | 5946 | | | | 5947 | | | | 5948 |Update | | | 5949 |---------------------------->| 5950 | | | | 5951 | | | | 5952 |UpdateAns| | | 5953 |<----------------------------| 5954 | | | | 5955 | | | | 5956 | | | | 5957 | | | | 5959 12. Security Considerations 5961 12.1. Overview 5963 RELOAD provides a generic storage service, albeit one designed to be 5964 useful for P2PSIP. In this section we discuss security issues that 5965 are likely to be relevant to any usage of RELOAD. More background 5966 information can be found in [RFC5765]. 5968 In any Overlay Instance, any given user depends on a number of peers 5969 with which they have no well-defined relationship except that they 5970 are fellow members of the Overlay Instance. In practice, these other 5971 nodes may be friendly, lazy, curious, or outright malicious. No 5972 security system can provide complete protection in an environment 5973 where most nodes are malicious. The goal of security in RELOAD is to 5974 provide strong security guarantees of some properties even in the 5975 face of a large number of malicious nodes and to allow the overlay to 5976 function correctly in the face of a modest number of malicious nodes. 5978 P2PSIP deployments require the ability to authenticate both peers and 5979 resources (users) without the active presence of a trusted entity in 5980 the system. We describe two mechanisms. The first mechanism is 5981 based on public key certificates and is suitable for general 5982 deployments. The second is an admission control mechanism based on 5983 an overlay-wide shared symmetric key. 5985 12.2. Attacks on P2P Overlays 5987 The two basic functions provided by overlay nodes are storage and 5988 routing: some node is responsible for storing a peer's data and for 5989 allowing a third peer to fetch this stored data. Other nodes are 5990 responsible for routing messages to and from the storing nodes. Each 5991 of these issues is covered in the following sections. 5993 P2P overlays are subject to attacks by subversive nodes that may 5994 attempt to disrupt routing, corrupt or remove user registrations, or 5995 eavesdrop on signaling. The certificate-based security algorithms we 5996 describe in this specification are intended to protect overlay 5997 routing and user registration information in RELOAD messages. 5999 To protect the signaling from attackers pretending to be valid peers 6000 (or peers other than themselves), the first requirement is to ensure 6001 that all messages are received from authorized members of the 6002 overlay. For this reason, RELOAD transports all messages over a 6003 secure channel (TLS and DTLS are defined in this document) which 6004 provides message integrity and authentication of the directly 6005 communicating peer. In addition, messages and data are digitally 6006 signed with the sender's private key, providing end-to-end security 6007 for communications. 6009 12.3. Certificate-based Security 6011 This specification stores users' registrations and possibly other 6012 data in an overlay network. This requires a solution to securing 6013 this data as well as securing, as well as possible, the routing in 6014 the overlay. Both types of security are based on requiring that 6015 every entity in the system (whether user or peer) authenticate 6016 cryptographically using an asymmetric key pair tied to a certificate. 6018 When a user enrolls in the Overlay Instance, they request or are 6019 assigned a unique name, such as "alice@dht.example.net". These names 6020 are unique and are meant to be chosen and used by humans much like a 6021 SIP Address of Record (AOR) or an email address. The user is also 6022 assigned one or more Node-IDs by the central enrollment authority. 6023 Both the name and the Node-ID are placed in the certificate, along 6024 with the user's public key. 6026 Each certificate enables an entity to act in two sorts of roles: 6028 o As a user, storing data at specific Resource-IDs in the Overlay 6029 Instance corresponding to the user name. 6030 o As a overlay peer with the Node-ID(s) listed in the certificate. 6032 Note that since only users of this Overlay Instance need to validate 6033 a certificate, this usage does not require a global PKI. Instead, 6034 certificates are signed by a central enrollment authority which acts 6035 as the certificate authority for the Overlay Instance. This 6036 authority signs each peer's certificate. Because each peer possesses 6037 the CA's certificate (which they receive on enrollment) they can 6038 verify the certificates of the other entities in the overlay without 6039 further communication. Because the certificates contain the user/ 6040 peer's public key, communications from the user/peer can be verified 6041 in turn. 6043 If self-signed certificates are used, then the security provided is 6044 significantly decreased, since attackers can mount Sybil attacks. In 6045 addition, attackers cannot trust the user names in certificates 6046 (though they can trust the Node-IDs because they are 6047 cryptographically verifiable). This scheme may be appropriate for 6048 some small deployments, such as a small office or an ad hoc overlay 6049 set up among participants in a meeting where all hosts on the network 6050 are trusted. Some additional security can be provided by using the 6051 shared secret admission control scheme as well. 6053 Because all stored data is signed by the owner of the data the 6054 storing peer can verify that the storer is authorized to perform a 6055 store at that Resource-ID and also allow any consumer of the data to 6056 verify the provenance and integrity of the data when it retrieves it. 6058 Note that RELOAD does not itself provide a revocation/status 6059 mechanism (though certificates may of course include OCSP responder 6060 information). Thus, certificate lifetimes should be chosen to 6061 balance the compromise window versus the cost of certificate renewal. 6062 Because RELOAD is already designed to operate in the face of some 6063 fraction of malicious peers, this form of compromise is not fatal. 6065 All implementations MUST implement certificate-based security. 6067 12.4. Shared-Secret Security 6069 RELOAD also supports a shared secret admission control scheme that 6070 relies on a single key that is shared among all members of the 6071 overlay. It is appropriate for small groups that wish to form a 6072 private network without complexity. In shared secret mode, all the 6073 peers share a single symmetric key which is used to key TLS-PSK 6074 [RFC4279] or TLS-SRP [RFC5054] mode. A peer which does not know the 6075 key cannot form TLS connections with any other peer and therefore 6076 cannot join the overlay. 6078 One natural approach to a shared-secret scheme is to use a user- 6079 entered password as the key. The difficulty with this is that in 6080 TLS-PSK mode, such keys are very susceptible to dictionary attacks. 6082 If passwords are used as the source of shared-keys, then TLS-SRP is a 6083 superior choice because it is not subject to dictionary attacks. 6085 12.5. Storage Security 6087 When certificate-based security is used in RELOAD, any given 6088 Resource-ID/Kind-ID pair is bound to some small set of certificates. 6089 In order to write data, the writer must prove possession of the 6090 private key for one of those certificates. Moreover, all data is 6091 stored, signed with the same private key that was used to authorize 6092 the storage. This set of rules makes questions of authorization and 6093 data integrity - which have historically been thorny for overlays - 6094 relatively simple. 6096 12.5.1. Authorization 6098 When a client wants to store some value, it first digitally signs the 6099 value with its own private key. It then sends a Store request that 6100 contains both the value and the signature towards the storing peer 6101 (which is defined by the Resource Name construction algorithm for 6102 that particular kind of value). 6104 When the storing peer receives the request, it must determine whether 6105 the storing client is authorized to store at this Resource-ID/Kind-ID 6106 pair. Determining this requires comparing the user's identity to the 6107 requirements of the access control model (see Section 6.3). If it 6108 satisfies those requirements the user is authorized to write, pending 6109 quota checks as described in the next section. 6111 For example, consider the certificate with the following properties: 6113 User name: alice@dht.example.com 6114 Node-ID: 013456789abcdef 6115 Serial: 1234 6117 If Alice wishes to Store a value of the "SIP Location" kind, the 6118 Resource Name will be the SIP AOR "sip:alice@dht.example.com". The 6119 Resource-ID will be determined by hashing the Resource Name. Because 6120 SIP Location uses the USER-NODE-MATCH policy, it first verifies that 6121 the user name in the certificate hashes to the requested Resource-ID. 6122 It then verifies that the Node-Id in the certificate matches the 6123 dictionary key being used for the store. If both of these checks 6124 succeed, the Store is authorized. Note that because the access 6125 control model is different for different kinds, the exact set of 6126 checks will vary. 6128 12.5.2. Distributed Quota 6130 Being a peer in an Overlay Instance carries with it the 6131 responsibility to store data for a given region of the Overlay 6132 Instance. However, allowing clients to store unlimited amounts of 6133 data would create unacceptable burdens on peers and would also enable 6134 trivial denial of service attacks. RELOAD addresses this issue by 6135 requiring configurations to define maximum sizes for each kind of 6136 stored data. Attempts to store values exceeding this size MUST be 6137 rejected (if peers are inconsistent about this, then strange 6138 artifacts will happen when the zone of responsibility shifts and a 6139 different peer becomes responsible for overlarge data). Because each 6140 Resource-ID/Kind-ID pair is bound to a small set of certificates, 6141 these size restrictions also create a distributed quota mechanism, 6142 with the quotas administered by the central configuration server. 6144 Allowing different kinds of data to have different size restrictions 6145 allows new usages the flexibility to define limits that fit their 6146 needs without requiring all usages to have expansive limits. 6148 12.5.3. Correctness 6150 Because each stored value is signed, it is trivial for any retrieving 6151 peer to verify the integrity of the stored value. Some more care 6152 needs to be taken to prevent version rollback attacks. Rollback 6153 attacks on storage are prevented by the use of store times and 6154 lifetime values in each store. A lifetime represents the latest time 6155 at which the data is valid and thus limits (though does not 6156 completely prevent) the ability of the storing node to perform a 6157 rollback attack on retrievers. In order to prevent a rollback attack 6158 at the time of the Store request, we require that storage times be 6159 monotonically increasing. Storing peers MUST reject Store requests 6160 with storage times smaller than or equal to those they are currently 6161 storing. In addition, a fetching node which receives a data value 6162 with a storage time older than the result of the previous fetch knows 6163 a rollback has occurred. 6165 12.5.4. Residual Attacks 6167 The mechanisms described here provides a high degree of security, but 6168 some attacks remain possible. Most simply, it is possible for 6169 storing nodes to refuse to store a value (i.e., reject any request). 6170 In addition, a storing node can deny knowledge of values which it has 6171 previously accepted. To some extent these attacks can be ameliorated 6172 by attempting to store to/retrieve from replicas, but a retrieving 6173 client does not know whether it should try this or not, since there 6174 is a cost to doing so. 6176 The certificate-based authentication scheme prevents a single peer 6177 from being able to forge data owned by other peers. Furthermore, 6178 although a subversive peer can refuse to return data resources for 6179 which it is responsible, it cannot return forged data because it 6180 cannot provide authentication for such registrations. Therefore 6181 parallel searches for redundant registrations can mitigate most of 6182 the effects of a compromised peer. The ultimate reliability of such 6183 an overlay is a statistical question based on the replication factor 6184 and the percentage of compromised peers. 6186 In addition, when a kind is multivalued (e.g., an array data model), 6187 the storing node can return only some subset of the values, thus 6188 biasing its responses. This can be countered by using single values 6189 rather than sets, but that makes coordination between multiple 6190 storing agents much more difficult. This is a trade off that must be 6191 made when designing any usage. 6193 12.6. Routing Security 6195 Because the storage security system guarantees (within limits) the 6196 integrity of the stored data, routing security focuses on stopping 6197 the attacker from performing a DOS attack that misroutes requests in 6198 the overlay. There are a few obvious observations to make about 6199 this. First, it is easy to ensure that an attacker is at least a 6200 valid peer in the Overlay Instance. Second, this is a DOS attack 6201 only. Third, if a large percentage of the peers on the Overlay 6202 Instance are controlled by the attacker, it is probably impossible to 6203 perfectly secure against this. 6205 12.6.1. Background 6207 In general, attacks on DHT routing are mounted by the attacker 6208 arranging to route traffic through one or two nodes it controls. In 6209 the Eclipse attack [Eclipse] the attacker tampers with messages to 6210 and from nodes for which it is on-path with respect to a given victim 6211 node. This allows it to pretend to be all the nodes that are 6212 reachable through it. In the Sybil attack [Sybil], the attacker 6213 registers a large number of nodes and is therefore able to capture a 6214 large amount of the traffic through the DHT. 6216 Both the Eclipse and Sybil attacks require the attacker to be able to 6217 exercise control over her Node-IDs. The Sybil attack requires the 6218 creation of a large number of peers. The Eclipse attack requires 6219 that the attacker be able to impersonate specific peers. In both 6220 cases, these attacks are limited by the use of centralized, 6221 certificate-based admission control. 6223 12.6.2. Admissions Control 6225 Admission to a RELOAD Overlay Instance is controlled by requiring 6226 that each peer have a certificate containing its Node-Id. The 6227 requirement to have a certificate is enforced by using certificate- 6228 based mutual authentication on each connection. (Note: the 6229 following only applies when self-signed certificates are not used.) 6230 Whenever a peer connects to another peer, each side automatically 6231 checks that the other has a suitable certificate. These Node-Ids are 6232 randomly assigned by the central enrollment server. This has two 6233 benefits: 6235 o It allows the enrollment server to limit the number of Node-IDs 6236 issued to any individual user. 6237 o It prevents the attacker from choosing specific Node-Ids. 6239 The first property allows protection against Sybil attacks (provided 6240 the enrollment server uses strict rate limiting policies). The 6241 second property deters but does not completely prevent Eclipse 6242 attacks. Because an Eclipse attacker must impersonate peers on the 6243 other side of the attacker, he must have a certificate for suitable 6244 Node-Ids, which requires him to repeatedly query the enrollment 6245 server for new certificates, which will match only by chance. From 6246 the attacker's perspective, the difficulty is that if he only has a 6247 small number of certificates, the region of the Overlay Instance he 6248 is impersonating appears to be very sparsely populated by comparison 6249 to the victim's local region. 6251 12.6.3. Peer Identification and Authentication 6253 In general, whenever a peer engages in overlay activity that might 6254 affect the routing table it must establish its identity. This 6255 happens in two ways. First, whenever a peer establishes a direct 6256 connection to another peer it authenticates via certificate-based 6257 mutual authentication. All messages between peers are sent over this 6258 protected channel and therefore the peers can verify the data origin 6259 of the last hop peer for requests and responses without further 6260 cryptography. 6262 In some situations, however, it is desirable to be able to establish 6263 the identity of a peer with whom one is not directly connected. The 6264 most natural case is when a peer Updates its state. At this point, 6265 other peers may need to update their view of the overlay structure, 6266 but they need to verify that the Update message came from the actual 6267 peer rather than from an attacker. To prevent this, all overlay 6268 routing messages are signed by the peer that generated them. 6270 Replay is typically prevented for messages that impact the topology 6271 of the overlay by having the information come directly, or be 6272 verified by, the nodes that claimed to have generated the update. 6273 Data storage replay detection is done by signing time of the node 6274 that generated the signature on the store request thus providing a 6275 time based replay protection but the time synchronization is only 6276 needed between peers that can write to the same location. 6278 12.6.4. Protecting the Signaling 6280 The goal here is to stop an attacker from knowing who is signaling 6281 what to whom. An attacker is unlikely to be able to observe the 6282 activities of a specific individual given the randomization of IDs 6283 and routing based on the present peers discussed above. Furthermore, 6284 because messages can be routed using only the header information, the 6285 actual body of the RELOAD message can be encrypted during 6286 transmission. 6288 There are two lines of defense here. The first is the use of TLS or 6289 DTLS for each communications link between peers. This provides 6290 protection against attackers who are not members of the overlay. The 6291 second line of defense is to digitally sign each message. This 6292 prevents adversarial peers from modifying messages in flight, even if 6293 they are on the routing path. 6295 12.6.5. Residual Attacks 6297 The routing security mechanisms in RELOAD are designed to contain 6298 rather than eliminate attacks on routing. It is still possible for 6299 an attacker to mount a variety of attacks. In particular, if an 6300 attacker is able to take up a position on the overlay routing between 6301 A and B it can make it appear as if B does not exist or is 6302 disconnected. It can also advertise false network metrics in an 6303 attempt to reroute traffic. However, these are primarily DOS 6304 attacks. 6306 The certificate-based security scheme secures the namespace, but if 6307 an individual peer is compromised or if an attacker obtains a 6308 certificate from the CA, then a number of subversive peers can still 6309 appear in the overlay. While these peers cannot falsify responses to 6310 resource queries, they can respond with error messages, effecting a 6311 DoS attack on the resource registration. They can also subvert 6312 routing to other compromised peers. To defend against such attacks, 6313 a resource search must still consist of parallel searches for 6314 replicated registrations. 6316 13. IANA Considerations 6318 This section contains the new code points registered by this 6319 document. [NOTE TO IANA/RFC-EDITOR: Please replace RFC-AAAA with 6320 the RFC number for this specification in the following list.] 6322 13.1. Well-Known URI Registration 6324 IANA will make the following "Well Known URI" registration as 6325 described in [RFC5785]: 6327 [[Note to RFC Editor - this paragraph can be removed before 6328 publication. ]] A review request was sent to 6329 wellknown-uri-review@ietf.org on October 12, 2010. 6331 +----------------------------+----------------------+ 6332 | URI suffix: | p2psip-enroll | 6333 | Change controller: | IETF | 6334 | Specification document(s): | [RFC-AAAA] | 6335 | Related information: | None | 6336 +----------------------------+----------------------+ 6338 13.2. Port Registrations 6340 [[Note to RFC Editor - this paragraph can be removed before 6341 publication. ]] IANA has already allocated a TCP port for the main 6342 peer to peer protocol. This port has the name p2p-sip and the port 6343 number of 6084. IANA needs to update this registration to be defined 6344 for UDP as well as TCP. 6346 IANA will make the following port registration: 6348 +------------------------------+------------------------------------+ 6349 | Registration Technical | Cullen Jennings | 6350 | Contact | | 6351 | Registration Owner | IETF | 6352 | Transport Protocol | TCP & UDP | 6353 | Port Number | 6084 | 6354 | Service Name | p2psip-enroll | 6355 | Description | Peer to Peer Infrastructure | 6356 | | Enrollment | 6357 | Reference | [RFC-AAAA] | 6358 +------------------------------+------------------------------------+ 6360 13.3. Overlay Algorithm Types 6362 IANA SHALL create a "RELOAD Overlay Algorithm Type" Registry. 6363 Entries in this registry are strings denoting the names of overlay 6364 algorithms. The registration policy for this registry is RFC 5226 6365 IETF Review. The initial contents of this registry are: 6367 +----------------+----------+ 6368 | Algorithm Name | RFC | 6369 +----------------+----------+ 6370 | CHORD-RELOAD | RFC-AAAA | 6371 +----------------+----------+ 6373 13.4. Access Control Policies 6375 IANA SHALL create a "RELOAD Access Control Policy" Registry. Entries 6376 in this registry are strings denoting access control policies, as 6377 described in Section 6.3. New entries in this registry SHALL be 6378 registered via RFC 5226 Standards Action. The initial contents of 6379 this registry are: 6381 +-----------------+----------+ 6382 | Access Policy | RFC | 6383 +-----------------+----------+ 6384 | USER-MATCH | RFC-AAAA | 6385 | NODE-MATCH | RFC-AAAA | 6386 | USER-NODE-MATCH | RFC-AAAA | 6387 | NODE-MULTIPLE | RFC-AAAA | 6388 +-----------------+----------+ 6390 13.5. Application-ID 6392 IANA SHALL create a "RELOAD Application-ID" Registry. Entries in 6393 this registry are 16-bit integers denoting application kinds. Code 6394 points in the range 0x0001 to 0x7fff SHALL be registered via RFC 5226 6395 Standards Action. Code points in the range 0x8000 to 0xf000 SHALL be 6396 registered via RFC 5226 Expert Review. Code points in the range 6397 0xf001 to 0xfffe are reserved for private use. The initial contents 6398 of this registry are: 6400 +-------------+----------------+-------------------------------+ 6401 | Application | Application-ID | Specification | 6402 +-------------+----------------+-------------------------------+ 6403 | INVALID | 0 | RFC-AAAA | 6404 | SIP | 5060 | Reserved for use by SIP Usage | 6405 | SIP | 5061 | Reserved for use by SIP Usage | 6406 | Reserved | 0xffff | RFC-AAAA | 6407 +-------------+----------------+-------------------------------+ 6409 13.6. Data Kind-ID 6411 IANA SHALL create a "RELOAD Data Kind-ID" Registry. Entries in this 6412 registry are 32-bit integers denoting data kinds, as described in 6413 Section 4.2. Code points in the range 0x00000001 to 0x7fffffff SHALL 6414 be registered via RFC 5226 Standards Action. Code points in the 6415 range 0x8000000 to 0xf0000000 SHALL be registered via RFC 5226 Expert 6416 Review. Code points in the range 0xf0000001 to 0xfffffffe are 6417 reserved for private use via the kind description mechanism described 6418 in Section 10. The initial contents of this registry are: 6420 +---------------------+------------+----------+ 6421 | Kind | Kind-ID | RFC | 6422 +---------------------+------------+----------+ 6423 | INVALID | 0 | RFC-AAAA | 6424 | TURN_SERVICE | 2 | RFC-AAAA | 6425 | CERTIFICATE_BY_NODE | 3 | RFC-AAAA | 6426 | CERTIFICATE_BY_USER | 16 | RFC-AAAA | 6427 | Reserved | 0x7fffffff | RFC-AAAA | 6428 | Reserved | 0xfffffffe | RFC-AAAA | 6429 +---------------------+------------+----------+ 6431 13.7. Data Model 6433 IANA SHALL create a "RELOAD Data Model" Registry. Entries in this 6434 registry denoting data models, as described in Section 6.2. Code 6435 points in this registry SHALL be registered via RFC 5226 Standards 6436 Action. The initial contents of this registry are: 6438 +------------+----------+ 6439 | Data Model | RFC | 6440 +------------+----------+ 6441 | INVALID | RFC-AAAA | 6442 | SINGLE | RFC-AAAA | 6443 | ARRAY | RFC-AAAA | 6444 | DICTIONARY | RFC-AAAA | 6445 | RESERVED | RFC-AAAA | 6446 +------------+----------+ 6448 13.8. Message Codes 6450 IANA SHALL create a "RELOAD Message Code" Registry. Entries in this 6451 registry are 16-bit integers denoting method codes as described in 6452 Section 5.3.3. These codes SHALL be registered via RFC 5226 6453 Standards Action. The initial contents of this registry are: 6455 +---------------------------------+----------------+----------+ 6456 | Message Code Name | Code Value | RFC | 6457 +---------------------------------+----------------+----------+ 6458 | invalid | 0 | RFC-AAAA | 6459 | probe_req | 1 | RFC-AAAA | 6460 | probe_ans | 2 | RFC-AAAA | 6461 | attach_req | 3 | RFC-AAAA | 6462 | attach_ans | 4 | RFC-AAAA | 6463 | unused | 5 | | 6464 | unused | 6 | | 6465 | store_req | 7 | RFC-AAAA | 6466 | store_ans | 8 | RFC-AAAA | 6467 | fetch_req | 9 | RFC-AAAA | 6468 | fetch_ans | 10 | RFC-AAAA | 6469 | unused (was remove_req) | 11 | RFC-AAAA | 6470 | unused (was remove_ans) | 12 | RFC-AAAA | 6471 | find_req | 13 | RFC-AAAA | 6472 | find_ans | 14 | RFC-AAAA | 6473 | join_req | 15 | RFC-AAAA | 6474 | join_ans | 16 | RFC-AAAA | 6475 | leave_req | 17 | RFC-AAAA | 6476 | leave_ans | 18 | RFC-AAAA | 6477 | update_req | 19 | RFC-AAAA | 6478 | update_ans | 20 | RFC-AAAA | 6479 | route_query_req | 21 | RFC-AAAA | 6480 | route_query_ans | 22 | RFC-AAAA | 6481 | ping_req | 23 | RFC-AAAA | 6482 | ping_ans | 24 | RFC-AAAA | 6483 | stat_req | 25 | RFC-AAAA | 6484 | stat_ans | 26 | RFC-AAAA | 6485 | unused (was attachlite_req) | 27 | RFC-AAAA | 6486 | unused (was attachlite_ans) | 28 | RFC-AAAA | 6487 | app_attach_req | 29 | RFC-AAAA | 6488 | app_attach_ans | 30 | RFC-AAAA | 6489 | unused (was app_attachlite_req) | 31 | RFC-AAAA | 6490 | unused (was app_attachlite_ans) | 32 | RFC-AAAA | 6491 | config_update_req | 33 | RFC-AAAA | 6492 | config_update_ans | 34 | RFC-AAAA | 6493 | reserved | 0x8000..0xfffe | RFC-AAAA | 6494 | error | 0xffff | RFC-AAAA | 6495 +---------------------------------+----------------+----------+ 6497 13.9. Error Codes 6499 IANA SHALL create a "RELOAD Error Code" Registry. Entries in this 6500 registry are 16-bit integers denoting error codes. New entries SHALL 6501 be defined via RFC 5226 Standards Action. The initial contents of 6502 this registry are: 6504 +-------------------------------------+----------------+----------+ 6505 | Error Code Name | Code Value | RFC | 6506 +-------------------------------------+----------------+----------+ 6507 | invalid | 0 | RFC-AAAA | 6508 | Unused | 1 | RFC-AAAA | 6509 | Error_Forbidden | 2 | RFC-AAAA | 6510 | Error_Not_Found | 3 | RFC-AAAA | 6511 | Error_Request_Timeout | 4 | RFC-AAAA | 6512 | Error_Generation_Counter_Too_Low | 5 | RFC-AAAA | 6513 | Error_Incompatible_with_Overlay | 6 | RFC-AAAA | 6514 | Error_Unsupported_Forwarding_Option | 7 | RFC-AAAA | 6515 | Error_Data_Too_Large | 8 | RFC-AAAA | 6516 | Error_Data_Too_Old | 9 | RFC-AAAA | 6517 | Error_TTL_Exceeded | 10 | RFC-AAAA | 6518 | Error_Message_Too_Large | 11 | RFC-AAAA | 6519 | Error_Unknown_Kind | 12 | RFC-AAAA | 6520 | Error_Unknown_Extension | 13 | RFC-AAAA | 6521 | Error_Response_Too_Large | 14 | RFC-AAAA | 6522 | Error_Config_Too_Old | 15 | RFC-AAAA | 6523 | Error_Config_Too_New | 16 | RFC-AAAA | 6524 | Error_In_Progress | 17 | RFC-AAAA | 6525 | reserved | 0x8000..0xfffe | RFC-AAAA | 6526 +-------------------------------------+----------------+----------+ 6528 13.10. Overlay Link Types 6530 IANA shall create a "RELOAD Overlay Link." New entries SHALL be 6531 defined via RFC 5226 Standards Action. This registry SHALL be 6532 initially populated with the following values: 6534 +--------------------+------+---------------+ 6535 | Protocol | Code | Specification | 6536 +--------------------+------+---------------+ 6537 | reserved | 0 | RFC-AAAA | 6538 | DTLS-UDP-SR | 1 | RFC-AAAA | 6539 | DTLS-UDP-SR-NO-ICE | 3 | RFC-AAAA | 6540 | TLS-TCP-FH-NO-ICE | 4 | RFC-AAAA | 6541 | reserved | 255 | RFC-AAAA | 6542 +--------------------+------+---------------+ 6544 13.11. Overlay Link Protocols 6546 IANA shall create an "Overlay Link Protocol Registry". Entries in 6547 this registry SHALL be defined via RFC 5226 Standards Action. This 6548 registry SHALL be initially populated with the following value: 6549 "TLS". 6551 13.12. Forwarding Options 6553 IANA shall create a "Forwarding Option Registry". Entries in this 6554 registry between 1 and 127 SHALL be defined via RFC 5226 Standards 6555 Action. Entries in this registry between 128 and 254 SHALL be 6556 defined via RFC 5226 Specification Required. This registry SHALL be 6557 initially populated with the following values: 6559 +-------------------+------+---------------+ 6560 | Forwarding Option | Code | Specification | 6561 +-------------------+------+---------------+ 6562 | invalid | 0 | RFC-AAAA | 6563 | reserved | 255 | RFC-AAAA | 6564 +-------------------+------+---------------+ 6566 13.13. Probe Information Types 6568 IANA shall create a "RELOAD Probe Information Type Registry". 6569 Entries in this registry SHALL be defined via RFC 5226 Standards 6570 Action. This registry SHALL be initially populated with the 6571 following values: 6573 +-----------------+------+---------------+ 6574 | Probe Option | Code | Specification | 6575 +-----------------+------+---------------+ 6576 | invalid | 0 | RFC-AAAA | 6577 | responsible_set | 1 | RFC-AAAA | 6578 | num_resources | 2 | RFC-AAAA | 6579 | uptime | 3 | RFC-AAAA | 6580 | reserved | 255 | RFC-AAAA | 6581 +-----------------+------+---------------+ 6583 13.14. Message Extensions 6585 IANA shall create a "RELOAD Extensions Registry". Entries in this 6586 registry SHALL be defined via RFC 5226 Specification Required. This 6587 registry SHALL be initially populated with the following values: 6589 +-----------------+--------+---------------+ 6590 | Extensions Name | Code | Specification | 6591 +-----------------+--------+---------------+ 6592 | invalid | 0 | RFC-AAAA | 6593 | reserved | 0xFFFF | RFC-AAAA | 6594 +-----------------+--------+---------------+ 6596 13.15. reload URI Scheme 6598 This section describes the scheme for a reload URI, which can be used 6599 to refer to either: 6601 o A peer. 6602 o A resource inside a peer. 6604 The reload URI is defined using a subset of the URI schema specified 6605 in Appendix A of RFC 3986 [RFC3986] and the associated URI Guidelines 6606 [RFC4395] per the following ABNF syntax: 6608 RELOAD-URI = "reload://" destination "@" overlay "/" 6609 [specifier] 6611 destination = 1 * HEXDIG 6612 overlay = reg-name 6613 specifier = 1*HEXDIG 6615 The definitions of these productions are as follows: 6617 destination: a hex-encoded Destination List object (i.e., multiple 6618 concatenated Destination objects with no length prefix prior to 6619 the object as a whole.) 6621 overlay: the name of the overlay. 6623 specifier : a hex-encoded StoredDataSpecifier indicating the data 6624 element. 6626 If no specifier is present then this URI addresses the peer which can 6627 be reached via the indicated destination list at the indicated 6628 overlay name. If a specifier is present, then the URI addresses the 6629 data value. 6631 13.15.1. URI Registration 6633 [[ Note to RFC Editor - please remove this paragraph before 6634 publication. ]] A review request was sent to uri-review@ietf.org on 6635 Oct 7, 2010. 6637 The following summarizes the information necessary to register the 6638 reload URI. 6640 URI Scheme Name: reload 6641 Status: permanent 6642 URI Scheme Syntax: see Section 13.15 of RFC-AAAA 6643 URI Scheme Semantics: The reload URI is intended to be used as a 6644 reference to a RELOAD peer or resource. 6645 Encoding Considerations: The reload URI is not intended to be human- 6646 readable text, so it is encoded entirely in US-ASCII. 6647 Applications/protocols that use this URI scheme: The RELOAD protocol 6648 described in RFC-AAAA. 6649 Interoperability considerations: See RFC-AAAA. 6650 Security considerations: See RFC-AAAA 6651 Contact: Cullen Jennings 6652 Author/Change controller: IESG 6653 References: RFC-AAAA 6655 13.16. Media Type Registration 6657 [[ Note to RFC Editor - please remove this paragraph before 6658 publication. ]] A review request was sent to ietf-types@iana.org on 6659 May 27, 2011. 6661 Type name: application 6663 Subtype name: p2p-overlay+xml 6665 Required parameters: none 6667 Optional parameters: none 6669 Encoding considerations: Must be binary encoded. 6671 Security considerations: This media type is typically not used to 6672 transport information that typically needs to be kept confidential 6673 however there are cases where it is integrity of the information is 6674 important. For these cases using a digital signature is RECOMMENDED. 6675 One way of doing this is specified in RFC-AAAA. In the case when the 6676 media includes a "shared-secret" element, then the contents of the 6677 file need to be kept confidential or else anyone that can see the 6678 shared-secret and effect the RELOAD overlay network. 6680 Interoperability considerations: No knows interoperability 6681 consideration beyond those identified for application/xml in 6682 [RFC3023]. 6684 Published specification: RFC-AAAA 6686 Applications that use this media type: The type is used to configure 6687 the peer to peer overlay networks defined in RFC-AAAA. 6689 Additional information: The syntax for this media type is specified 6690 in Section 10.1 of RFC-AAAA. The contents MUST be valid XML 6691 compliant with the relax NG grammar specified in RFC-AAAA and use the 6692 UTF-8[RFC3629] character encoding. 6694 Magic number(s): none 6696 File extension(s): relo 6698 Macintosh file type code(s): none 6700 Person & email address to contact for further information: Cullen 6701 Jennings 6703 Intended usage: COMMON 6705 Restrictions on usage: None 6707 Author: Cullen Jennings 6709 Change controller: IESG 6711 14. Acknowledgments 6713 This specification is a merge of the "REsource LOcation And Discovery 6714 (RELOAD)" draft by David A. Bryan, Marcia Zangrilli and Bruce B. 6715 Lowekamp, the "Address Settlement by Peer to Peer" draft by Cullen 6716 Jennings, Jonathan Rosenberg, and Eric Rescorla, the "Security 6717 Extensions for RELOAD" draft by Bruce B. Lowekamp and James Deverick, 6718 the "A Chord-based DHT for Resource Lookup in P2PSIP" by Marcia 6719 Zangrilli and David A. Bryan, and the Peer-to-Peer Protocol (P2PP) 6720 draft by Salman A. Baset, Henning Schulzrinne, and Marcin 6721 Matuszewski. Thanks to the authors of RFC 5389 for text included 6722 from that. Vidya Narayanan provided many comments and improvements. 6724 The ideas and text for the Chord specific extension data to the Leave 6725 mechanisms was provided by Jouni Maenpaa, Gonzalo Camarillo, and Jani 6726 Hautakorpi. 6728 Thanks to the many people who contributed including Ted Hardie, 6729 Michael Chen, Dan York, Das Saumitra, Lyndsay Campbell, Brian Rosen, 6730 David Bryan, Dave Craig, and Julian Cain. Extensive working last 6731 call comments were provided by: Jouni Maenpaa, Roni Even, Gonzalo 6732 Camarillo, Ari Keranen, John Buford, Michael Chen, Frederic-Philippe 6733 Met, and David Bryan. Special thanks to Marc Petit-Huguenin who 6734 provied an amazing amount to detailed review. 6736 15. References 6738 15.1. Normative References 6740 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 6741 Requirement Levels", BCP 14, RFC 2119, March 1997. 6743 [RFC2585] Housley, R. and P. Hoffman, "Internet X.509 Public Key 6744 Infrastructure Operational Protocols: FTP and HTTP", 6745 RFC 2585, May 1999. 6747 [RFC2818] Rescorla, E., "HTTP Over TLS", RFC 2818, May 2000. 6749 [RFC3023] Murata, M., St. Laurent, S., and D. Kohn, "XML Media 6750 Types", RFC 3023, January 2001. 6752 [RFC3174] Eastlake, D. and P. Jones, "US Secure Hash Algorithm 1 6753 (SHA1)", RFC 3174, September 2001. 6755 [RFC3447] Jonsson, J. and B. Kaliski, "Public-Key Cryptography 6756 Standards (PKCS) #1: RSA Cryptography Specifications 6757 Version 2.1", RFC 3447, February 2003. 6759 [RFC3629] Yergeau, F., "UTF-8, a transformation format of ISO 6760 10646", STD 63, RFC 3629, November 2003. 6762 [RFC3986] Berners-Lee, T., Fielding, R., and L. Masinter, "Uniform 6763 Resource Identifier (URI): Generic Syntax", STD 66, 6764 RFC 3986, January 2005. 6766 [RFC4279] Eronen, P. and H. Tschofenig, "Pre-Shared Key Ciphersuites 6767 for Transport Layer Security (TLS)", RFC 4279, 6768 December 2005. 6770 [RFC4347] Rescorla, E. and N. Modadugu, "Datagram Transport Layer 6771 Security", RFC 4347, April 2006. 6773 [RFC4395] Hansen, T., Hardie, T., and L. Masinter, "Guidelines and 6774 Registration Procedures for New URI Schemes", BCP 35, 6775 RFC 4395, February 2006. 6777 [RFC4648] Josefsson, S., "The Base16, Base32, and Base64 Data 6778 Encodings", RFC 4648, October 2006. 6780 [RFC5245] Rosenberg, J., "Interactive Connectivity Establishment 6781 (ICE): A Protocol for Network Address Translator (NAT) 6782 Traversal for Offer/Answer Protocols", RFC 5245, 6783 April 2010. 6785 [RFC5246] Dierks, T. and E. Rescorla, "The Transport Layer Security 6786 (TLS) Protocol Version 1.2", RFC 5246, August 2008. 6788 [RFC5272] Schaad, J. and M. Myers, "Certificate Management over CMS 6789 (CMC)", RFC 5272, June 2008. 6791 [RFC5273] Schaad, J. and M. Myers, "Certificate Management over CMS 6792 (CMC): Transport Protocols", RFC 5273, June 2008. 6794 [RFC5348] Floyd, S., Handley, M., Padhye, J., and J. Widmer, "TCP 6795 Friendly Rate Control (TFRC): Protocol Specification", 6796 RFC 5348, September 2008. 6798 [RFC5389] Rosenberg, J., Mahy, R., Matthews, P., and D. Wing, 6799 "Session Traversal Utilities for NAT (STUN)", RFC 5389, 6800 October 2008. 6802 [RFC5766] Mahy, R., Matthews, P., and J. Rosenberg, "Traversal Using 6803 Relays around NAT (TURN): Relay Extensions to Session 6804 Traversal Utilities for NAT (STUN)", RFC 5766, April 2010. 6806 [RFC5952] Kawamura, S. and M. Kawashima, "A Recommendation for IPv6 6807 Address Text Representation", RFC 5952, August 2010. 6809 [RFC6091] Mavrogiannopoulos, N. and D. Gillmor, "Using OpenPGP Keys 6810 for Transport Layer Security (TLS) Authentication", 6811 RFC 6091, February 2011. 6813 [RFC6234] Eastlake, D. and T. Hansen, "US Secure Hash Algorithms 6814 (SHA and SHA-based HMAC and HKDF)", RFC 6234, May 2011. 6816 [RFC6298] Paxson, V., Allman, M., Chu, J., and M. Sargent, 6817 "Computing TCP's Retransmission Timer", RFC 6298, 6818 June 2011. 6820 [w3c-xml-namespaces] 6821 Bray, T., Hollander, D., Layman, A., Tobin, R., and Henry 6822 S. , "Namespaces in XML 1.0 (Third Edition)". 6824 15.2. Informative References 6826 [Chord] Stoica, I., Morris, R., Liben-Nowell, D., Karger, D., 6827 Kaashoek, M., Dabek, F., and H. Balakrishnan, "Chord: A 6828 Scalable Peer-to-peer Lookup Protocol for Internet 6829 Applications", IEEE/ACM Transactions on Networking Volume 6830 11, Issue 1, 17-32, Feb 2003. 6832 [Eclipse] Singh, A., Ngan, T., Druschel, T., and D. Wallach, 6833 "Eclipse Attacks on Overlay Networks: Threats and 6834 Defenses", INFOCOM 2006, April 2006. 6836 [I-D.ietf-hip-reload-instance] 6837 Keranen, A., Camarillo, G., and J. Maenpaa, "Host Identity 6838 Protocol-Based Overlay Networking Environment (HIP BONE) 6839 Instance Specification for REsource LOcation And Discovery 6840 (RELOAD)", draft-ietf-hip-reload-instance-03 (work in 6841 progress), January 2011. 6843 [I-D.ietf-mmusic-ice-tcp] 6844 Rosenberg, J., Keranen, A., Lowekamp, B., and A. Roach, 6845 "TCP Candidates with Interactive Connectivity 6846 Establishment (ICE)", draft-ietf-mmusic-ice-tcp-13 (work 6847 in progress), April 2011. 6849 [I-D.ietf-p2psip-self-tuning] 6850 Maenpaa, J., Camarillo, G., and J. Hautakorpi, "A Self- 6851 tuning Distributed Hash Table (DHT) for REsource LOcation 6852 And Discovery (RELOAD)", draft-ietf-p2psip-self-tuning-04 6853 (work in progress), July 2011. 6855 [I-D.ietf-p2psip-service-discovery] 6856 Maenpaa, J. and G. Camarillo, "Service Discovery Usage for 6857 REsource LOcation And Discovery (RELOAD)", 6858 draft-ietf-p2psip-service-discovery-03 (work in progress), 6859 July 2011. 6861 [I-D.ietf-p2psip-sip] 6862 Jennings, C., Lowekamp, B., Rescorla, E., Baset, S., and 6863 H. Schulzrinne, "A SIP Usage for RELOAD", 6864 draft-ietf-p2psip-sip-06 (work in progress), July 2011. 6866 [I-D.jiang-p2psip-relay] 6867 Jiang, X., Zong, N., Even, R., and Y. Zhang, "An extension 6868 to RELOAD to support Direct Response and Relay Peer 6869 routing", draft-jiang-p2psip-relay-05 (work in progress), 6870 March 2011. 6872 [I-D.pascual-p2psip-clients] 6873 Pascual, V., Matuszewski, M., Shim, E., Zhang, H., and S. 6874 Yongchao, "P2PSIP Clients", 6875 draft-pascual-p2psip-clients-01 (work in progress), 6876 February 2008. 6878 [RFC1122] Braden, R., "Requirements for Internet Hosts - 6879 Communication Layers", STD 3, RFC 1122, October 1989. 6881 [RFC2311] Dusse, S., Hoffman, P., Ramsdell, B., Lundblade, L., and 6882 L. Repka, "S/MIME Version 2 Message Specification", 6883 RFC 2311, March 1998. 6885 [RFC4086] Eastlake, D., Schiller, J., and S. Crocker, "Randomness 6886 Requirements for Security", BCP 106, RFC 4086, June 2005. 6888 [RFC4145] Yon, D. and G. Camarillo, "TCP-Based Media Transport in 6889 the Session Description Protocol (SDP)", RFC 4145, 6890 September 2005. 6892 [RFC4787] Audet, F. and C. Jennings, "Network Address Translation 6893 (NAT) Behavioral Requirements for Unicast UDP", BCP 127, 6894 RFC 4787, January 2007. 6896 [RFC4828] Floyd, S. and E. Kohler, "TCP Friendly Rate Control 6897 (TFRC): The Small-Packet (SP) Variant", RFC 4828, 6898 April 2007. 6900 [RFC5054] Taylor, D., Wu, T., Mavrogiannopoulos, N., and T. Perrin, 6901 "Using the Secure Remote Password (SRP) Protocol for TLS 6902 Authentication", RFC 5054, November 2007. 6904 [RFC5201] Moskowitz, R., Nikander, P., Jokela, P., and T. Henderson, 6905 "Host Identity Protocol", RFC 5201, April 2008. 6907 [RFC5280] Cooper, D., Santesson, S., Farrell, S., Boeyen, S., 6908 Housley, R., and W. Polk, "Internet X.509 Public Key 6909 Infrastructure Certificate and Certificate Revocation List 6910 (CRL) Profile", RFC 5280, May 2008. 6912 [RFC5694] Camarillo, G. and IAB, "Peer-to-Peer (P2P) Architecture: 6913 Definition, Taxonomies, Examples, and Applicability", 6914 RFC 5694, November 2009. 6916 [RFC5765] Schulzrinne, H., Marocco, E., and E. Ivov, "Security 6917 Issues and Solutions in Peer-to-Peer Systems for Realtime 6918 Communications", RFC 5765, February 2010. 6920 [RFC5785] Nottingham, M. and E. Hammer-Lahav, "Defining Well-Known 6921 Uniform Resource Identifiers (URIs)", RFC 5785, 6922 April 2010. 6924 [RFC6079] Camarillo, G., Nikander, P., Hautakorpi, J., Keranen, A., 6925 and A. Johnston, "HIP BONE: Host Identity Protocol (HIP) 6926 Based Overlay Networking Environment (BONE)", RFC 6079, 6927 January 2011. 6929 [Sybil] Douceur, J., "The Sybil Attack", IPTPS 02, March 2002. 6931 [UnixTime] 6932 Wikipedia, "Unix Time", . 6935 [bryan-design-hotp2p08] 6936 Bryan, D., Lowekamp, B., and M. Zangrilli, "The Design of 6937 a Versatile, Secure P2PSIP Communications Architecture for 6938 the Public Internet", Hot-P2P'08. 6940 [handling-churn-usenix04] 6941 Rhea, S., Geels, D., Roscoe, T., and J. Kubiatowicz, 6942 "Handling Churn in a DHT", In Proc. of the USENIX Annual 6943 Technical Conference June 2004 USENIX 2004. 6945 [lookups-churn-p2p06] 6946 Wu, D., Tian, Y., and K. Ng, "Analytical Study on 6947 Improving DHT Lookup Performance under Churn", IEEE 6948 P2P'06. 6950 [minimizing-churn-sigcomm06] 6951 Godfrey, P., Shenker, S., and I. Stoica, "Minimizing Churn 6952 in Distributed Systems", SIGCOMM 2006. 6954 [non-transitive-dhts-worlds05] 6955 Freedman, M., Lakshminarayanan, K., Rhea, S., and I. 6956 Stoica, "Non-Transitive Connectivity and DHTs", 6957 WORLDS'05. 6959 [opendht-sigcomm05] 6960 Rhea, S., Godfrey, B., Karp, B., Kubiatowicz, J., 6961 Ratnasamy, S., Shenker, S., Stoica, I., and H. Yu, 6962 "OpenDHT: A Public DHT and its Uses", SIGCOMM'05. 6964 [vulnerabilities-acsac04] 6965 Srivatsa, M. and L. Liu, "Vulnerabilities and Security 6966 Threats in Structured Peer-to-Peer Systems: A Quantitative 6967 Analysis", ACSAC 2004. 6969 Appendix A. Routing Alternatives 6971 Significant discussion has been focused on the selection of a routing 6972 algorithm for P2PSIP. This section discusses the motivations for 6973 selecting symmetric recursive routing for RELOAD and describes the 6974 extensions that would be required to support additional routing 6975 algorithms. 6977 A.1. Iterative vs Recursive 6979 Iterative routing has a number of advantages. It is easier to debug, 6980 consumes fewer resources on intermediate peers, and allows the 6981 querying peer to identify and route around misbehaving peers 6982 [non-transitive-dhts-worlds05]. However, in the presence of NATs, 6983 iterative routing is intolerably expensive because a new connection 6984 must be established for each hop (using ICE) [bryan-design-hotp2p08]. 6986 Iterative routing is supported through the RouteQuery mechanism and 6987 is primarily intended for debugging. It also allows the querying 6988 peer to evaluate the routing decisions made by the peers at each hop, 6989 consider alternatives, and perhaps detect at what point the 6990 forwarding path fails. 6992 A.2. Symmetric vs Forward response 6994 An alternative to the symmetric recursive routing method used by 6995 RELOAD is Forward-Only routing, where the response is routed to the 6996 requester as if it were a new message initiated by the responder (in 6997 the previous example, Z sends the response to A as if it were sending 6998 a request). Forward-only routing requires no state in either the 6999 message or intermediate peers. 7001 The drawback of forward-only routing is that it does not work when 7002 the overlay is unstable. For example, if A is in the process of 7003 joining the overlay and is sending a Join request to Z, it is not yet 7004 reachable via forward routing. Even if it is established in the 7005 overlay, if network failures produce temporary instability, A may not 7006 be reachable (and may be trying to stabilize its network connectivity 7007 via Attach messages). 7009 Furthermore, forward-only responses are less likely to reach the 7010 querying peer than symmetric recursive ones are, because the forward 7011 path is more likely to have a failed peer than is the request path 7012 (which was just tested to route the request) 7013 [non-transitive-dhts-worlds05]. 7015 An extension to RELOAD that supports forward-only routing but relies 7016 on symmetric responses as a fallback would be possible, but due to 7017 the complexities of determining when to use forward-only and when to 7018 fallback to symmetric, we have chosen not to include it as an option 7019 at this point. 7021 A.3. Direct Response 7023 Another routing option is Direct Response routing, in which the 7024 response is returned directly to the querying node. In the previous 7025 example, if A encodes its IP address in the request, then Z can 7026 simply deliver the response directly to A. In the absence of NATs or 7027 other connectivity issues, this is the optimal routing technique. 7029 The challenge of implementing direct response is the presence of 7030 NATs. There are a number of complexities that must be addressed. In 7031 this discussion, we will continue our assumption that A issued the 7032 request and Z is generating the response. 7034 o The IP address listed by A may be unreachable, either due to NAT 7035 or firewall rules. Therefore, a direct response technique must 7036 fallback to symmetric response [non-transitive-dhts-worlds05]. 7037 The hop-by-hop ACKs used by RELOAD allow Z to determine when A has 7038 received the message (and the TLS negotiation will provide earlier 7039 confirmation that A is reachable), but this fallback requires a 7040 timeout that will increase the response latency whenever A is not 7041 reachable from Z. 7042 o Whenever A is behind a NAT it will have multiple candidate IP 7043 addresses, each of which must be advertised to ensure 7044 connectivity; therefore Z will need to attempt multiple 7045 connections to deliver the response. 7046 o One (or all) of A's candidate addresses may route from Z to a 7047 different device on the Internet. In the worst case these nodes 7048 may actually be running RELOAD on the same port. Therefore, it is 7049 absolutely necessary to establish a secure connection to 7050 authenticate A before delivering the response. This step 7051 diminishes the efficiency of direct response because multiple 7052 roundtrips are required before the message can be delivered. 7053 o If A is behind a NAT and does not have a connection already 7054 established with Z, there are only two ways the direct response 7055 will work. The first is that A and Z both be behind the same NAT, 7056 in which case the NAT is not involved. In the more common case, 7057 when Z is outside A's NAT, the response will only be received if 7058 A's NAT implements endpoint-independent filtering. As the choice 7059 of filtering mode conflates application transparency with security 7060 [RFC4787], and no clear recommendation is available, the 7061 prevalence of this feature in future devices remains unclear. 7063 An extension to RELOAD that supports direct response routing but 7064 relies on symmetric responses as a fallback would be possible, but 7065 due to the complexities of determining when to use direct response 7066 and when to fallback to symmetric, and the reduced performance for 7067 responses to peers behind restrictive NATs, we have chosen not to 7068 include it as an option at this point. 7070 A.4. Relay Peers 7072 [I-D.jiang-p2psip-relay] has proposed implementing a form of direct 7073 response by having A identify a peer, Q, that will be directly 7074 reachable by any other peer. A uses Attach to establish a connection 7075 with Q and advertises Q's IP address in the request sent to Z. Z 7076 sends the response to Q, which relays it to A. This then reduces the 7077 latency to two hops, plus Z negotiating a secure connection to Q. 7079 This technique relies on the relative population of nodes such as A 7080 that require relay peers and peers such as Q that are capable of 7081 serving as a relay peer. It also requires nodes to be able to 7082 identify which category they are in. This identification problem has 7083 turned out to be hard to solve and is still an open area of 7084 exploration. 7086 An extension to RELOAD that supports relay peers is possible, but due 7087 to the complexities of implementing such an alternative, we have not 7088 added such a feature to RELOAD at this point. 7090 A concept similar to relay peers, essentially choosing a relay peer 7091 at random, has previously been suggested to solve problems of 7092 pairwise non-transitivity [non-transitive-dhts-worlds05], but 7093 deterministic filtering provided by NATs makes random relay peers no 7094 more likely to work than the responding peer. 7096 A.5. Symmetric Route Stability 7098 A common concern about symmetric recursive routing has been that one 7099 or more peers along the request path may fail before the response is 7100 received. The significance of this problem essentially depends on 7101 the response latency of the overlay. An overlay that produces slow 7102 responses will be vulnerable to churn, whereas responses that are 7103 delivered very quickly are vulnerable only to failures that occur 7104 over that small interval. 7106 The other aspect of this issue is whether the request itself can be 7107 successfully delivered. Assuming typical connection maintenance 7108 intervals, the time period between the last maintenance and the 7109 request being sent will be orders of magnitude greater than the delay 7110 between the request being forwarded and the response being received. 7111 Therefore, if the path was stable enough to be available to route the 7112 request, it is almost certainly going to remain available to route 7113 the response. 7115 An overlay that is unstable enough to suffer this type of failure 7116 frequently is unlikely to be able to support reliable functionality 7117 regardless of the routing mechanism. However, regardless of the 7118 stability of the return path, studies show that in the event of high 7119 churn, iterative routing is a better solution to ensure request 7120 completion [lookups-churn-p2p06] [non-transitive-dhts-worlds05] 7122 Finally, because RELOAD retries the end-to-end request, that retry 7123 will address the issues of churn that remain. 7125 Appendix B. Why Clients? 7127 There are a wide variety of reasons a node may act as a client rather 7128 than as a peer [I-D.pascual-p2psip-clients]. This section outlines 7129 some of those scenarios and how the client's behavior changes based 7130 on its capabilities. 7132 B.1. Why Not Only Peers? 7134 For a number of reasons, a particular node may be forced to act as a 7135 client even though it is willing to act as a peer. These include: 7137 o The node does not have appropriate network connectivity, typically 7138 because it has a low-bandwidth network connection. 7139 o The node may not have sufficient resources, such as computing 7140 power, storage space, or battery power. 7141 o The overlay algorithm may dictate specific requirements for peer 7142 selection. These may include participating in the overlay to 7143 determine trustworthiness; controlling the number of peers in the 7144 overlay to reduce overly-long routing paths; or ensuring minimum 7145 application uptime before a node can join as a peer. 7147 The ultimate criteria for a node to become a peer are determined by 7148 the overlay algorithm and specific deployment. A node acting as a 7149 client that has a full implementation of RELOAD and the appropriate 7150 overlay algorithm is capable of locating its responsible peer in the 7151 overlay and using Attach to establish a direct connection to that 7152 peer. In that way, it may elect to be reachable under either of the 7153 routing approaches listed above. Particularly for overlay algorithms 7154 that elect nodes to serve as peers based on trustworthiness or 7155 population, the overlay algorithm may require such a client to locate 7156 itself at a particular place in the overlay. 7158 B.2. Clients as Application-Level Agents 7160 SIP defines an extensive protocol for registration and security 7161 between a client and its registrar/proxy server(s). Any SIP device 7162 can act as a client of a RELOAD-based P2PSIP overlay if it contacts a 7163 peer that implements the server-side functionality required by the 7164 SIP protocol. In this case, the peer would be acting as if it were 7165 the user's peer, and would need the appropriate credentials for that 7166 user. 7168 Application-level support for clients is defined by a usage. A usage 7169 offering support for application-level clients should specify how the 7170 security of the system is maintained when the data is moved between 7171 the application and RELOAD layers. 7173 Appendix C. Change Log 7175 C.1. Changes since draft-ietf-p2psip-reload-13 7177 Note to RFC Editor: Please remove this section. 7179 o Fixed finger formula. 7180 o Support for getting a certificate with multiple Node-IDs. 7181 o Added configuration signer. 7182 o Added way to specify mandatory extensions in configuration file. 7184 Authors' Addresses 7186 Cullen Jennings 7187 Cisco 7188 170 West Tasman Drive 7189 MS: SJC-21/2 7190 San Jose, CA 95134 7191 USA 7193 Phone: +1 408 421-9990 7194 Email: fluffy@cisco.com 7196 Bruce B. Lowekamp (editor) 7197 Skype 7198 Palo Alto, CA 7199 USA 7201 Email: bbl@lowekamp.net 7202 Eric Rescorla 7203 RTFM, Inc. 7204 2064 Edgewood Drive 7205 Palo Alto, CA 94303 7206 USA 7208 Phone: +1 650 678 2350 7209 Email: ekr@rtfm.com 7211 Salman A. Baset 7212 Columbia University 7213 1214 Amsterdam Avenue 7214 New York, NY 7215 USA 7217 Email: salman@cs.columbia.edu 7219 Henning Schulzrinne 7220 Columbia University 7221 1214 Amsterdam Avenue 7222 New York, NY 7223 USA 7225 Email: hgs@cs.columbia.edu