idnits 2.17.1 draft-ietf-nsis-ntlp-05.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- ** It looks like you're using RFC 3978 boilerplate. You should update this to the boilerplate described in the IETF Trust License Policy document (see https://trustee.ietf.org/license-info), which is required now. -- Found old boilerplate from RFC 3978, Section 5.1.a on line 18. -- Found old boilerplate from RFC 3978, Section 5.5 on line 3929. -- Found old boilerplate from RFC 3979, Section 5, paragraph 1 on line 3906. -- Found old boilerplate from RFC 3979, Section 5, paragraph 2 on line 3913. -- Found old boilerplate from RFC 3979, Section 5, paragraph 3 on line 3919. ** The document seems to lack an RFC 3978 Section 5.1 IPR Disclosure Acknowledgement. ** This document has an original RFC 3978 Section 5.4 Copyright Line, instead of the newer IETF Trust Copyright according to RFC 4748. ** This document has an original RFC 3978 Section 5.5 Disclaimer, instead of the newer disclaimer which includes the IETF Trust according to RFC 4748. ** The document uses RFC 3667 boilerplate or RFC 3978-like boilerplate instead of verbatim RFC 3978 boilerplate. After 6 May 2005, submission of drafts without verbatim RFC 3978 boilerplate is not accepted. The following non-3978 patterns matched text found in the document. That text should be removed or replaced: This document is an Internet-Draft and is subject to all provisions of Section 3 of RFC 3667. By submitting this Internet-Draft, each author represents that any applicable patent or other IPR claims of which he or she is aware have been or will be disclosed, and any of which he or she becomes aware will be disclosed, in accordance with Section 6 of BCP 79. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- == No 'Intended status' indicated for this document; assuming Proposed Standard Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- ** There is 1 instance of too long lines in the document, the longest one being 1 character in excess of 72. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the RFC 3978 Section 5.4 Copyright Line does not match the current year == Line 3218 has weird spacing: '... | ti on ...' -- The document seems to lack a disclaimer for pre-RFC5378 work, but may have content which was first submitted before 10 November 2008. If you have contacted all the original authors and they are all willing to grant the BCP78 rights to the IETF Trust, then this is fine, and you can ignore this comment. If not, you may need to add the pre-RFC5378 disclaimer. (See the Legal Provisions document at https://trustee.ietf.org/license-info for more information.) -- The document date (February 21, 2005) is 7004 days in the past. Is this intentional? Checking references for intended status: Proposed Standard ---------------------------------------------------------------------------- (See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) -- Looks like a reference, but probably isn't: 'Data' on line 216 -- Looks like a reference, but probably isn't: 'Flow' on line 233 == Unused Reference: '1' is defined on line 3041, but no explicit reference was found in the text == Unused Reference: '4' is defined on line 3049, but no explicit reference was found in the text == Unused Reference: '10' is defined on line 3072, but no explicit reference was found in the text == Unused Reference: '16' is defined on line 3091, but no explicit reference was found in the text == Unused Reference: '18' is defined on line 3099, but no explicit reference was found in the text == Unused Reference: '22' is defined on line 3113, but no explicit reference was found in the text ** Obsolete normative reference: RFC 2234 (ref. '3') (Obsoleted by RFC 4234) ** Obsolete normative reference: RFC 2765 (ref. '5') (Obsoleted by RFC 6145) ** Obsolete normative reference: RFC 2960 (ref. '6') (Obsoleted by RFC 4960) == Outdated reference: A later version (-13) exists of draft-ietf-dccp-spec-09 == Outdated reference: A later version (-07) exists of draft-ietf-ipngwg-icmp-v3-06 -- Possible downref: Normative reference to a draft: ref. '8' -- Obsolete informational reference (is this intentional?): RFC 2409 (ref. '10') (Obsoleted by RFC 4306) -- Obsolete informational reference (is this intentional?): RFC 2766 (ref. '12') (Obsoleted by RFC 4966) -- Obsolete informational reference (is this intentional?): RFC 3068 (ref. '14') (Obsoleted by RFC 7526) -- Obsolete informational reference (is this intentional?): RFC 3489 (ref. '19') (Obsoleted by RFC 5389) == Outdated reference: A later version (-25) exists of draft-ietf-nsis-nslp-natfw-04 == Outdated reference: A later version (-18) exists of draft-ietf-nsis-qos-nslp-05 == Outdated reference: A later version (-07) exists of draft-ietf-v6ops-mech-v2-06 == Outdated reference: A later version (-22) exists of draft-ietf-secsh-architecture-21 == Outdated reference: A later version (-10) exists of draft-ietf-hip-base-01 == Outdated reference: A later version (-03) exists of draft-ietf-mip6-ro-sec-02 == Outdated reference: A later version (-04) exists of draft-bound-dstm-exp-02 == Outdated reference: A later version (-02) exists of draft-stiemerling-nsis-natfw-mrm-01 Summary: 9 errors (**), 0 flaws (~~), 19 warnings (==), 14 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Next Steps in Signaling H. Schulzrinne 3 Internet-Draft Columbia U. 4 Expires: August 25, 2005 R. Hancock 5 Siemens/RMR 6 February 21, 2005 8 GIMPS: General Internet Messaging Protocol for Signaling 9 draft-ietf-nsis-ntlp-05 11 Status of this Memo 13 This document is an Internet-Draft and is subject to all provisions 14 of Section 3 of RFC 3667. By submitting this Internet-Draft, each 15 author represents that any applicable patent or other IPR claims of 16 which he or she is aware have been or will be disclosed, and any of 17 which he or she become aware will be disclosed, in accordance with 18 RFC 3668. 20 Internet-Drafts are working documents of the Internet Engineering 21 Task Force (IETF), its areas, and its working groups. Note that 22 other groups may also distribute working documents as 23 Internet-Drafts. 25 Internet-Drafts are draft documents valid for a maximum of six months 26 and may be updated, replaced, or obsoleted by other documents at any 27 time. It is inappropriate to use Internet-Drafts as reference 28 material or to cite them other than as "work in progress." 30 The list of current Internet-Drafts can be accessed at 31 http://www.ietf.org/ietf/1id-abstracts.txt. 33 The list of Internet-Draft Shadow Directories can be accessed at 34 http://www.ietf.org/shadow.html. 36 This Internet-Draft will expire on August 25, 2005. 38 Copyright Notice 40 Copyright (C) The Internet Society (2005). 42 Abstract 44 This document specifies protocol stacks for the routing and transport 45 of per-flow signaling messages along the path taken by that flow 46 through the network. The design uses existing transport and security 47 protocols under a common messaging layer, the General Internet 48 Messaging Protocol for Signaling (GIMPS), which provides a universal 49 service for diverse signaling applications. GIMPS does not handle 50 signaling application state itself, but manages its own internal 51 state and the configuration of the underlying transport and security 52 protocols to enable the transfer of messages in both directions along 53 the flow path. The combination of GIMPS and the lower layer 54 transport and security protocols provides a solution for the base 55 protocol component of the "Next Steps in Signaling" framework. 57 Table of Contents 59 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 4 60 1.1 Restrictions on Scope . . . . . . . . . . . . . . . . . . 5 61 2. Requirements Notation and Terminology . . . . . . . . . . . 6 62 3. Design Overview . . . . . . . . . . . . . . . . . . . . . . 8 63 3.1 Overall Design Approach . . . . . . . . . . . . . . . . . 8 64 3.2 Example of Operation . . . . . . . . . . . . . . . . . . . 10 65 4. GIMPS Processing Overview . . . . . . . . . . . . . . . . . 14 66 4.1 GIMPS Service Interface . . . . . . . . . . . . . . . . . 14 67 4.2 GIMPS State . . . . . . . . . . . . . . . . . . . . . . . 16 68 4.3 Basic Message Processing . . . . . . . . . . . . . . . . . 18 69 4.4 Routing State and Messaging Association Maintenance . . . 22 70 5. Message Formats and Transport . . . . . . . . . . . . . . . 28 71 5.1 GIMPS Messages . . . . . . . . . . . . . . . . . . . . . . 28 72 5.2 Information Elements . . . . . . . . . . . . . . . . . . . 29 73 5.3 Datagram Mode Transport . . . . . . . . . . . . . . . . . 33 74 5.4 Connection Mode Transport . . . . . . . . . . . . . . . . 38 75 5.5 Messaging Association Negotiation . . . . . . . . . . . . 40 76 6. Advanced Protocol Features . . . . . . . . . . . . . . . . . 43 77 6.1 Route Changes and Local Repair . . . . . . . . . . . . . . 43 78 6.2 Policy-Based Forwarding and Flow Wildcarding . . . . . . . 49 79 6.3 NAT Traversal . . . . . . . . . . . . . . . . . . . . . . 49 80 6.4 Interaction with IP Tunnelling . . . . . . . . . . . . . . 51 81 6.5 IPv4-IPv6 Transition and Interworking . . . . . . . . . . 52 82 7. Security Considerations . . . . . . . . . . . . . . . . . . 54 83 7.1 Message Confidentiality and Integrity . . . . . . . . . . 54 84 7.2 Peer Node Authentication . . . . . . . . . . . . . . . . . 55 85 7.3 Routing State Integrity . . . . . . . . . . . . . . . . . 55 86 7.4 Denial of Service Prevention . . . . . . . . . . . . . . . 57 87 8. IANA Considerations . . . . . . . . . . . . . . . . . . . . 59 88 9. Open Issues . . . . . . . . . . . . . . . . . . . . . . . . 61 89 9.1 Additional Discovery Mechanisms . . . . . . . . . . . . . 61 90 9.2 Alternative Message Routing Requirements . . . . . . . . . 61 91 9.3 Message Format Issues . . . . . . . . . . . . . . . . . . 62 92 10. Change History . . . . . . . . . . . . . . . . . . . . . . . 64 93 10.1 Changes In Version -05 . . . . . . . . . . . . . . . . . 64 94 10.2 Changes In Version -04 . . . . . . . . . . . . . . . . . 65 95 10.3 Changes In Version -03 . . . . . . . . . . . . . . . . . 66 96 10.4 Changes In Version -02 . . . . . . . . . . . . . . . . . 67 97 10.5 Changes In Version -01 . . . . . . . . . . . . . . . . . 68 98 11. References . . . . . . . . . . . . . . . . . . . . . . . . . 71 99 11.1 Normative References . . . . . . . . . . . . . . . . . . 71 100 11.2 Informative References . . . . . . . . . . . . . . . . . 71 101 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . 73 102 A. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 74 103 B. Example Message Routing State Table . . . . . . . . . . . . 75 104 C. Bit-Level Formats . . . . . . . . . . . . . . . . . . . . . 77 105 C.1 General NSIS Formatting Guidelines . . . . . . . . . . . . 77 106 C.2 The GIMPS Common Header . . . . . . . . . . . . . . . . . 78 107 C.3 General Object Characteristics . . . . . . . . . . . . . . 78 108 C.4 GIMPS Specific TLV Objects . . . . . . . . . . . . . . . . 79 109 C.5 Generic NSIS TLV Objects . . . . . . . . . . . . . . . . . 85 110 D. API between GIMPS and NSLP . . . . . . . . . . . . . . . . . 87 111 D.1 SendMessage . . . . . . . . . . . . . . . . . . . . . . . 87 112 D.2 RecvMessage . . . . . . . . . . . . . . . . . . . . . . . 89 113 D.3 MessageStatus . . . . . . . . . . . . . . . . . . . . . . 90 114 D.4 NetworkNotification . . . . . . . . . . . . . . . . . . . 90 115 D.5 SetStateLifetime . . . . . . . . . . . . . . . . . . . . . 90 116 D.6 InvalidateRoutingState . . . . . . . . . . . . . . . . . . 91 117 Intellectual Property and Copyright Statements . . . . . . . 92 119 1. Introduction 121 Signaling involves the manipulation of state held in network 122 elements. 'Manipulation' could mean setting up, modifying and 123 tearing down state; or it could simply mean the monitoring of state 124 which is managed by other mechanisms. 126 This specification concentrates specifically on the case of 127 "path-coupled" signaling, which involves network elements which are 128 located on the path taken by a particular data flow, possibly 129 including but not limited to the flow endpoints. Indeed, there are 130 almost always more than two participants in a path-coupled-signaling 131 session, although there is no need for every router on the path to 132 participate. Path-coupled signaling thus excludes end-to-end 133 higher-layer application signaling (except as a degenerate case) such 134 as ISUP (telephony signaling for Signaling System #7) messages being 135 transported by SCTP between two nodes. 137 In the context of path-coupled signaling, examples of state 138 management include network resource allocation (for "resource 139 reservation"), firewall configuration, and state used in active 140 networking; examples of state monitoring are the discovery of 141 instantaneous path properties (such as available bandwidth, or 142 cumulative queuing delay). Each of these different uses of 143 path-coupled signaling is referred to as a signaling application. 145 Every signaling application requires a set of state management rules, 146 as well as protocol support to exchange messages along the data path. 147 Several aspects of this support are common to all or a large number 148 of signaling applications, and hence should be developed as a common 149 protocol. The framework given in [20] provides a rationale for a 150 function split between the common and application specific protocols, 151 and gives outline requirements for the former, the 'NSIS Transport 152 Layer Protocol' (NTLP). 154 This specification provides a concrete solution for the NTLP. It is 155 based on the use of existing transport and security protocols under a 156 common messaging layer, the General Internet Messaging Protocol for 157 Signaling (GIMPS). Different signaling applications may make use of 158 different services provided by GIMPS. However, GIMPS does not handle 159 signaling application state itself; in that crucial respect, it 160 differs from application signaling protocols such as the control 161 component of FTP, SIP and RTSP. Instead, GIMPS manages its own 162 internal state and the configuration of the underlying transport and 163 security protocols to ensure the transfer of signaling messages on 164 behalf of signaling applications in both directions along the flow 165 path. 167 1.1 Restrictions on Scope 169 This section briefly lists some important restrictions on GIMPS 170 applicability and functionality. In some cases, these are implicit 171 consequences of the functionality split developed in the framework; 172 in others, they are restrictions on the types of scenario in which 173 GIMPS can operate correctly. 175 Flow splitting: In some cases, e.g. where packet-level load sharing 176 has been implemented, the path taken by a single flow in the 177 network may not be well defined. If this is the case, GIMPS 178 cannot route signaling meaningfully. (In some circumstances, 179 GIMPS can detect this condition, but even this cannot be 180 guaranteed.) 182 Multicast: GIMPS does not handle multicast flows. This includes 183 'classical' IP multicast and any of the 'small group multicast' 184 schemes recently proposed. 186 2. Requirements Notation and Terminology 188 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 189 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 190 document are to be interpreted as described in [2]. 192 The terminology used in this specification is fully defined in this 193 section. The basic entities relevant at the GIMPS level are shown in 194 Figure 1. 196 Source GIMPS (adjacent) peer nodes Destination 198 IP address IP addresses = Signaling IP address 199 = Flow Source/Destination Addresses = Flow 200 Source (depending on signaling direction) Destination 201 Address | | Address 202 V V 203 +--------+ +------+ Data Flow +------+ +--------+ 204 | Flow |-----------|------|-------------|------|-------->| Flow | 205 | Sender | | | | | |Receiver| 206 +--------+ |GIMPS |============>|GIMPS | +--------+ 207 | Node |<============| Node | 208 +------+ Signaling +------+ 209 GN1 Flow GN2 211 >>>>>>>>>>>>>>>>> = Downstream direction 212 <<<<<<<<<<<<<<<<< = Upstream direction 214 Figure 1: Basic Terminology 216 [Data] Flow: A set of packets identified by some fixed combination of 217 header fields. Flows are unidirectional (a bidirectional 218 communication is considered a pair of unidirectional flows). 220 Session: A single application layer flow of information for which 221 some network control state information is to be manipulated or 222 monitored. IP mobility may cause the mapping between sessions and 223 flows to change, and IP multihoming may mean there is more than 224 one flow for a given session. GIMPS implements the session 225 concept by allowing signaling applications to associate messages 226 with a Session Identifier; however, GIMPS does not place any 227 constraints on how this association should be done. 229 [Flow] Sender: The node in the network which is the source of the 230 packets in a flow. Could be a host, or a router (e.g. if the 231 flow is actually an aggregate). 233 [Flow] Receiver: The node in the network which is the sink for the 234 packets in a flow. 236 Downstream: In the same direction as the data flow. 238 Upstream: In the opposite direction to the data flow. 240 GIMPS Node: Any node along the data path supporting GIMPS (regardless 241 of what signaling applications it supports). 243 Adjacent peer: The next GIMPS node along the data path, in the 244 upstream or downstream direction. Whether two nodes are adjacent 245 is determined implicitly by the GIMPS peer discovery mechanisms; 246 it is possible for adjacencies to 'skip over' intermediate GIMPS 247 nodes if it can be determined that they have no interest in the 248 signaling messages being exchanged. 250 Datagram mode: A mode of sending GIMPS messages between nodes without 251 using any transport layer state or security protection. Datagram 252 mode uses UDP encapsulation, with IP addresses derived either from 253 the flow definition or previously discovered adjacency 254 information; the details depend on the direction of the message. 256 Connection mode: A mode of sending GIMPS messages directly between 257 nodes using point to point "messaging associations" (see below). 258 Connection mode allows the re-use of existing transport and 259 security protocols where such functionality is required. 261 Messaging association: A single connection between two explicitly 262 identified GIMPS adjacent peers, i.e. between a given signaling 263 source and destination address. A messaging association may use a 264 specific transport protocol and known ports. If security 265 protection is required, it may use a specific network layer 266 security association, or use a transport layer security 267 association internally. A messaging association is bidirectional; 268 signaling messages can be sent over it in either direction, and 269 can refer to flows of either direction. 271 Message Routing Method: Even in the path-coupled case, there can be 272 different alogorithms for discovering the route that signaling 273 messages should take. These are referred to as message routing 274 methods, and GIMPS supports alternatives within a common protocol 275 framework. See also Section 4.2.1. 277 Transfer Attributes: A description of the requirements which a 278 signaling application has for the delivery of a particular 279 message; for example, whether the message should be delivered 280 reliably. See Section 4.1.2. 282 3. Design Overview 284 3.1 Overall Design Approach 286 The generic requirements identified in the NSIS framework [20] for 287 transport of path-coupled signaling messages are essentially 288 two-fold: 290 "Routing": Determine how to reach the adjacent signaling node along 291 each direction of the data path (the GIMPS peer), and if necessary 292 explicitly establish the identity of that peer; 294 "Transport": Deliver the signaling information to that peer. 296 To meet the routing requirement, one possibility is for the node to 297 use local routing state information to determine the identity of the 298 GIMPS peer explicitly. GIMPS defines a 3-way handshake 299 (Query/Response/optional Confirm) which sets up the necessary routing 300 state between adjacent peers; the Query message is encapsulated in a 301 special way, depending on the message routing method, in order to 302 probe the network infrastructure so that the correct peer will 303 intercept it. If the routing state does not exist, it may be 304 possible for GIMPS to send a message anyway, with the same 305 encapsulation tricks as used for a Query. 307 Once the routing decision has been made, the node has to select a 308 mechanism for transport of the message to the peer. GIMPS divides 309 the transport problems into two categories, the easy and the 310 difficult ones. It handles the easy cases internally, and uses 311 well-understood reliable transport protocols for the harder cases. 312 Here, with details discussed later, "easy" messages are those that 313 are sized well below the lowest MTU along a path, are infrequent 314 enough not to cause concerns about congestion and flow control, and 315 do not need transport or network-layer security protection or 316 guaranteed delivery. 318 In [20] all of these routing and transport requirements are assigned 319 to a single notional protocol, the 'NSIS Transport Layer Protocol' 320 (NTLP). The strategy of splitting the transport problem leads to a 321 layered structure for the NTLP, as a specialised GIMPS 'messaging' 322 layer running over standard transport and security protocols, as 323 shown in Figure 2. This also shows GIMPS offering its services to 324 upper layers at an abstract interface, the GIMPS API, further 325 discussed in Section 4.1. 327 Internally, GIMPS has two modes of operation: 329 Datagram mode is used for small, infrequent messages with modest 330 delay constraints; it is also used at least for the Query message 331 of the 3-way handshake. 333 Connection mode is used for larger data objects or where fast state 334 setup in the face of packet loss is desirable, or where channel 335 security is required. 337 ^^ +-------------+ 338 || | Signaling | 339 NSIS +------------|Application 2| 340 Signaling | Signaling +-------------+ 341 Application |Application 1| | 342 Level +-------------+ | 343 || | | 344 VV | | 345 =========|===================|===== <-- GIMPS API 346 | | 347 ^^ +------------------------------------------------+ 348 || |+-----------------------+ +--------------+ | 349 || || GIMPS | | GIMPS State | | 350 || || Encapsulation |<<<>>>| Maintenance | | 351 || |+-----------------------+ +--------------+ | 352 || |GIMPS: Messaging Layer | 353 || +------------------------------------------------+ 354 NSIS | | | | 355 Transport ............................. 356 Level . Transport Layer Security . 357 ("NTLP") ............................. 358 || | | | | 359 || +----+ +----+ +----+ +----+ 360 || |UDP | |TCP | |SCTP| |DCCP|.... 361 || +----+ +----+ +----+ +----+ 362 || | | | | 363 || ............................. 364 || . IP Layer Security . 365 || ............................. 366 VV | | | | 367 =========================|=======|=======|=======|=============== 368 | | | | 369 +----------------------------------------------+ 370 | IP | 371 +----------------------------------------------+ 373 Figure 2: Protocol Stacks for Signaling Transport 375 Datagram mode uses UDP, as this is the only encapsulation which does 376 not require shared state to be established between the peers. The 377 connection mode can in principal use any stream or message-oriented 378 transport protocol; this specification currently defines the use of 379 TCP as the initial choice. It may employ specific network layer 380 security associations (such as IPsec), or an internal transport layer 381 security association (such as TLS). 383 It is possible to mix these two modes along a chain of nodes, without 384 coordination or manual configuration. This allows, for example, the 385 use of datagram mode at the edges of the network and connection mode 386 in the core of the network. Such combinations may make operation 387 more efficient for mobile endpoints, while allowing multiplexing of 388 signaling messages across shared security associations and transport 389 connections between core routers. 391 It must be understood that the routing and transport decisions made 392 by GIMPS are not totally independent. If the message transfer has 393 requirements that enforce the use of connection mode (e.g. the 394 message is so large that fragmentation is required), this can only be 395 used between explicitly identified nodes. In such cases, GIMPS must 396 carry out the 3-way handshake initially in datagram mode to identify 397 the peer and then set up the necessary transport connection if it 398 does not already exist. It must also be understood that the 399 signaling application does not make the datagram vs. connection mode 400 selection directly; rather, this decision is made by GIMPS on the 401 basis of the message characteristics and the transfer attributes 402 stated by the application. The distinction is not visible at the 403 GIMPS service interface. 405 In general, the state associated with connection mode messaging to a 406 particular peer (signaling destination address, protocol and port 407 numbers, internal protocol configuration and state information) is 408 referred to as a "messaging association". There may be any number of 409 messaging associations between two GIMPS peers (although the usual 410 case is 0 or 1), and they are set up and torn down by management 411 actions within GIMPS itself. 413 3.2 Example of Operation 415 This section presents an example of GIMPS usage in a relatively 416 simple (in particular, NAT-free) signaling scenario, to illustrate 417 its main features. 419 Consider the case of an RSVP-like signaling application which 420 allocates resources for a flow from sender to receiver. We will 421 consider how GIMPS transfers messages between two adjacent peers 422 along the path, GN1 and GN2 (see Figure 1). In this example, the 423 end-to-end exchange is initiated by the signaling application 424 instance in the sender; we take up the story at the point where the 425 first message is being processed (above the GIMPS layer) by the 426 signaling application in GN1. 428 1. The signaling application in GN1 determines that this message is 429 a simple description of resources that would be appropriate for 430 the flow. It determines that it has no special security or 431 transport requirements for the message, but simply that it should 432 be transferred to the next downstream signaling application peer 433 on the path that the flow will take. 435 2. The message payload is passed to the GIMPS layer in GN1, along 436 with a definition of the flow and description of the message 437 transfer attributes {downstream, unsecured, unreliable}. GIMPS 438 determines that this particular message does not require 439 fragmentation and that it has no knowledge of the next peer for 440 this flow and signaling application; however, it also determines 441 that this application is likely to require secured upstream and 442 downstream transport of large messages in the future. This 443 determination is a function of node-local policy, and some 444 options for how it may be communicated between NSLP and GIMPS 445 implementations within a node are indicated in Appendix D. 447 3. GN1 therefore constructs a GIMPS-Query message, which is a UDP 448 datagram carrying the signaling application payload and 449 additional payloads at the GIMPS level to be used to initiate the 450 setup of a messaging association. The Query is injected into the 451 network, addressed towards the flow destination and with a Router 452 Alert Option included. 454 4. The Query message passes through the network towards the flow 455 receiver, and is seen by each router in turn. GIMPS-unaware 456 routers will not recognise the RAO value and will forward the 457 message unchanged; GIMPS-aware routers which do not support the 458 signaling application in question will also forward the message 459 basically unchanged, although they may need to process more of 460 the message to decide this. 462 5. The message is intercepted at GN2. The GIMPS layer identifies 463 the message as relevant to a local signaling application, and 464 passes the signaling application payload and flow description 465 upwards to it. There, the signaling application in GN2 continues 466 to process this message as in GN1 (compare step 1), and this will 467 eventually result in the message reaching the flow receiver. 469 6. In parallel, the GIMPS instance in GN2 recognises, by the fact 470 that the message is a GIMPS-Query, that GN1 is attempting to 471 discover GN2 in order to set up a messaging association for 472 future signaling for the flow. There are two possible cases for 473 sending back the necessary GIMPS-Response: 475 A. GN1 and GN2 already have an appropriate association. GN2 476 simply records the identity of GN1 as its upstream peer for 477 that flow and signaling application, and sends a 478 GIMPS-Response back to GN1 over the association identifying 479 itself as the peer for this flow. 481 B. No messaging association exists. Again, GN2 records the 482 identity of GN1 as before, but sends a GIMPS-Response 483 upstream to GN1, identifying itself and agreeing to the 484 association setup. The protocol exchanges needed to complete 485 this will proceed in the background, controlled by GN1. 487 7. Eventually, another signaling application message works its way 488 upstream from the receiver to GN2. This message contains a 489 description of the actual resources requested, along with 490 authorisation and other security information. The signaling 491 application in GN2 passes this payload to the GIMPS level, along 492 with the flow definition and transfer attributes {upstream, 493 secured, reliable}. 495 8. The GIMPS layer in GN2 identifies the upstream peer for this flow 496 and signaling application as GN1, and determines that it has a 497 messaging association with the appropriate properties. The 498 message is queued on the association for transmission (this may 499 mean some delay if the negotiations begun in step 6.B have not 500 yet completed). 502 Further messages can be passed in each direction in the same way. 503 The GIMPS layer in each node can in parallel carry out maintenance 504 operations such as route change detection (this can be done by 505 sending additional GIMPS-Query messages, see Section 6.1 for more 506 details). 508 Note that when GIMPS messages are carried in connection mode, they 509 are treated just like any other traffic by intermediate routers 510 between the GIMPS peers. Indeed, it would be impossible for 511 intermediate routers to carry out any processing on the messages 512 without terminating the transport and security protocols used. In 513 connection mode, signaling messages are only ever delivered between 514 peers established in GIMPS-Query/Response exchanges. Any route 515 change is not detected until another GIMPS-Query/Response procedure 516 takes place; in the meantime, signaling messages are misdelivered. 517 GIMPS is responsible for prompt detection of route changes to 518 minimise the period during which this can take place. 520 It should be understood that several of these details of GIMPS 521 operations can be varied, either by local policy or according to 522 signaling application requirements, and they are also subject to 523 development and refinement as the protocol design proceeds. The 524 authoritative details are contained in the remainder of this 525 document. 527 4. GIMPS Processing Overview 529 This section defines the basic structure and operation of GIMPS. It 530 is divided into four parts. Section 4.1 describes the way in which 531 GIMPS interacts with (local) signaling applications in the form of an 532 abstract service interface. Section 4.2 describes the per-flow and 533 per-peer state that GIMPS maintains for the purpose of transferring 534 messages. Section 4.3 describes how messages are processed in the 535 case where any necessary messaging associations and associated 536 routing state already exist; this includes the simple scenario of 537 pure datagram mode operation, where no messaging associations are 538 necessary in the first place. Finally, Section 4.4 describes how 539 routing state is maintained and how messaging associations are 540 initiated and terminated. 542 4.1 GIMPS Service Interface 544 This section defines the service interface that GIMPS presents to 545 signaling applications in terms of abstract properties of the message 546 transfer. Note that the same service interface is presented at every 547 GIMPS node; however, applications may invoke it differently at 548 different nodes (e.g. depending on local policy). In addition, the 549 service interface is defined independently of any specific transport 550 protocol, or even the distinction between datagram and connection 551 mode. The initial version of this specification defines how to 552 support the service interface using a connection mode based on TCP; 553 if additional transport protocol support is added, this will support 554 the same interface and so be invisible to applications (except as a 555 possible performance improvement). A more detailed specification of 556 this service interface is given in Appendix D. 558 4.1.1 Message Handling 560 Fundamentally, GIMPS provides a simple message-by-message transfer 561 service for use by signaling applications: individual messages are 562 sent, and individual messages are received. Messages consist of an 563 opaque signaling application payload, and control information 564 expressing the application's requirements about how the message 565 should be routed. Additional message transfer attributes control the 566 specific transport and security properties that the signaling 567 application desires for the message. 569 The distinction between GIMPS connection and datagram modes is not 570 visible at the service interface. In addition, the invocation of 571 GIMPS functionality to handle fragmentation and reassembly, bundling 572 together of small messages (for efficiency), and congestion control 573 are not directly visible at the service interface; GIMPS will take 574 whatever action is necessary based on the properties of the messages 575 and local node state. 577 Messages for different sessions (i.e. with different Session IDs, 578 see Section 4.2.1) are treated entirely independently of each other 579 by GIMPS. Messages for the same session which are to be delivered 580 reliably (see below) to the same peer will be delivered in order. If 581 the receiving application delays reading these messages, this will 582 (eventually) cause a flow-control condition at the sending node. 584 4.1.2 Message Transfer Attributes 586 Message transfer attributes are used to define certain 587 performance-related aspects of message processing. The attributes 588 available are as follows: 590 Reliability: This attribute may be 'true' or 'false'. For the case 591 'true', messages will be delivered to the signaling application in 592 the peer exactly once or not at all; if there is a chance that the 593 message was not delivered, an error will be indicated to the local 594 signaling application identifying the routing information for the 595 message in question. For the case 'false', a message may be 596 delivered, once, several times or not at all, with no error 597 indications in any case. 599 Security: This attribute defines the security properties that the 600 signaling application requires for the message, including the type 601 of protection required, and what authenticated identities should 602 be used for the signaling source and destination. This 603 information maps onto the corresponding properties of the security 604 associations established between the peers in connection mode, It 605 can be specified explicitly by the signaling application, or 606 reported by GIMPS to the signaling application (either on 607 receiving a message, or just before sending a message but after 608 configuring or selecting the messaging association to be used for 609 it). Further details are discussed in Appendix D. 611 Local Processing: An NSLP may provide hints to GIMPS to enable more 612 efficient or appropriate processing. The NSLP may select a 613 priority from a range of locally defined values to influence the 614 sequence in which messages leave a node. Any priority mechanism 615 must respect the ordering requirements for reliable messages 616 within a session, and priority values are not carried in the 617 protocol or available at the signaling peer or intermediate nodes. 618 An NSLP may also indicate that reverse path routing state will not 619 be needed for this flow, to inhibit the node requesting its 620 downstream peer to create it. 622 4.2 GIMPS State 624 4.2.1 Message Routing State 626 For each flow, the GIMPS layer can maintain message routing state to 627 manage the processing of outgoing messages. This state is 628 conceptually organised into a table with the following structure. 630 The primary key (index) for the table is the combination of the 631 information about how the message is to be routed, the session being 632 signalled for, and the signaling application itself: 634 Message Routing Information (MRI): This defines the method to be used 635 to route the message, and any associated addressing information. 636 In the commonest case, the message routing method is to follow the 637 path that is being taken by the data flow, and the associated 638 addressing is the flow header N-tuple (i.e. the Flow-Identifier 639 of [20]). Other message routing methods are possible, as 640 described for example in [29]. 642 Signaling Application Identification (NSLPID): This is an IANA 643 assigned identifier of the signaling application which is 644 generating messages for this flow. The inclusion of this 645 identifier allows the routing state to be different for different 646 signaling applications (e.g. because of different adjacencies). 648 Session Identification (SID): This is a cryptographically random and 649 (probabilistically) globally unique identifier of the application 650 layer session that is using the flow. For a given flow, different 651 signaling applications may or may not use the same session 652 identifier. Often there will only be one flow for a given 653 session, but in mobility/multihoming scenarios there may be more 654 than one and they may be differently routed. 656 For a given MRI and NSLPID the message routing state should not be 657 SID-dependent. The SID is included in the key as a barrier to 658 routing state being corrupted by a malicious upstream node. 660 The state information for a given key consists of two items, namely 661 the information needed to send messages to the peers in each 662 direction respectively. In each case, the information could be an IP 663 address and UDP port, or a pointer to a valid messaging association, 664 either of which can be learned from a prior GIMPS handshake. 665 Additional information about the number of IP hops to the peer is 666 also stored in the table for each direction. An example of a routing 667 state table for a simple scenario is given in Appendix B. 669 It is also possible for the state information for either direction to 670 be null. There are several possible cases: 672 o The signaling application has indicated that no messages will 673 actually be sent in that direction. 675 o The node is a flow endpoint, so there can be no signaling peer in 676 one or other direction. 678 o The node can use other techniques to route the message. For 679 example, it could encapsulate it the same way as a Query message 680 and rely on the peer to intercept it. 682 In addition, the SID itself is not actually required for message 683 processing; in that case, no state information at all needs to be 684 stored in the table. 686 Both items of state have associated timers for how long the 687 identification can be considered accurate; when these timers expire, 688 the peer identification is purged if it has not been refreshed. 689 Message routing state is installed and refreshed by the exchange of 690 GIMPS-Query/Response messages as described in Section 4.4. For a 691 given flow, the GIMPS node which initiated the state setup is 692 responsible for scheduling a Query/Response exchange to refresh it, 693 and to allow its peer to do likewise. This should be done while 694 GIMPS determines the signaling application is still active. GIMPS 695 may opportunistically synchronise these 'internal' refresh operations 696 with those in the signaling application if it wishes. 698 Note also that the information is described as a table of flows, but 699 that there is no implied constraint on how the information is stored. 700 For example, in a network using pure destination address routing 701 (without load sharing or any form of policy-based forwarding), the 702 downstream peer information might be possible to store in an 703 aggregated form in the same manner as the IP forwarding table. In 704 addition, many of the per-flow entries may point to the same per-peer 705 state (e.g. the same messaging association) if the flows go through 706 the same adjacent peer. However, in general, and especially if GIMPS 707 peers are several IP hops away, there is no way to identify the 708 correct downstream peer for a flow and signaling application from the 709 local forwarding table using prefix matching, and the same applies 710 always to upstream peer state because of the possibility of 711 asymmetric routing: per-flow routing state has to be stored, just as 712 for RSVP [9]. 714 4.2.2 Messaging Association State 716 The per-flow message routing state is not the only state stored by 717 GIMPS. There is also the state required to manage the messaging 718 associations. Since these associations are typically per-peer rather 719 than per-flow, they are stored in a separate table, including the 720 following information: 722 o messages pending transmission while an association is being 723 established; 725 o an inactivity timer for how long the association has been idle. 727 In addition, per-association state is held in the messaging 728 association protocols themselves. However, the details of this state 729 are not directly visible to GIMPS, and they do not affect the rest of 730 the protocol description. 732 4.3 Basic Message Processing 734 This section describes how signaling application messages are 735 processed in the case where any necessary messaging associations and 736 routing state are already in place. The description is divided into 737 several parts. Firstly, message reception, local processing and 738 message transmission are described for the case where the node 739 handles the NSLPID in the message. Secondly, the case where the 740 message is forwarded directly in the IP or GIMPS layer (because there 741 is no matching signaling application on the node) is given. An 742 overview is given in Figure 3. 744 Note that the same messages are used both for maintaining internal 745 GIMPS state and carrying signaling application payloads. The state 746 maintenance takes place as a result of processing specific GIMPS 747 payloads in these messages. The processing of these payloads is the 748 subject of Section 4.4. 750 4.3.1 Message Reception 752 Messages can be received in connection or datagram mode, and from 753 upstream or downstream peers. 755 Reception in connection mode is simple: incoming packets undergo the 756 security and transport treatment associated with the messaging 757 association, and the messaging association provides complete messages 758 to the GIMPS layer for further processing. Unless the message is 759 protected by a query/response cookie exchange (see Section 4.4), the 760 routing state table is checked to ensure that this messaging 761 association is associated with the MRI/NSLPID combination. 763 Reception in datagram mode depends on the message type. 'Normal' 764 messages arrive UDP encapsulated and addressed directly to the 765 receiving signaling node, at an address and port learned during a 766 previous handshake. Each datagram contains a single complete message 767 which is passed to the GIMPS layer for further processing, just as in 768 the connection mode case. 770 +---------------------------------------------------------+ 771 | >> Signaling Application Processing >> | 772 | | 773 +--------^---------------------------------------V--------+ 774 ^ V 775 ^ NSLP Payloads V 776 ^ V 777 +--------^---------------------------------------V--------+ 778 | >> GIMPS >> | 779 | ^ ^ ^ Processing V V V | 780 +--x-----------N--Q---------------------Q--N-----------x--+ 781 x N Q Q N x 782 x N Q>>>>>>>>>>>>>>>>>>>>>Q N x 783 x N Q Bypass at Q N x 784 +--x-----+ +--N--Q--+ GIMPS level +--Q--N--+ +-----x--+ 785 | C-mode | | D-mode | | D-mode | | C-mode | 786 |Handling| |Handling| |Handling| |Handling| 787 +--x-----+ +--N--Q--+ +--Q--N--+ +-----x--+ 788 x N Q Q N x 789 x NNNNNN Q>>>>>>>>>>>>>>>>>>>>>Q NNNNNN x 790 x N Q Bypass at Q N x 791 +--x--N--+ +-----Q--+ router +--Q-----+ +--N--x--+ 792 |IP Host | | RAO | alert level | RAO | |IP Host | 793 |Handling| |Handling| |Handling| |Handling| 794 +--x--N--+ +-----Q--+ +--Q-----+ +--N--x--+ 795 x N Q Q N x 796 +--x--N-----------Q--+ +--Q-----------N--x--+ 797 | IP Layer | | IP Layer | 798 | (Receive Side) | | (Transmit Side) | 799 +--x--N-----------Q--+ +--Q-----------N--x--+ 800 x N Q Q N x 801 x N Q Q N x 802 x N Q Q N x 804 NNNNNNNNNNNNNN = 'Normal' datagram mode messages 805 QQQQQQQQQQQQQQ = Datagram mode messages which 806 are Queries or likewise encapsulated 807 xxxxxxxxxxxxxx = connection mode messages 808 RAO = Router Alert Option 810 Figure 3: Message Paths through a GIMPS Node 812 Where GIMPS is sending messages to be intercepted by the appropriate 813 peer rather than directly addressed to it (in particular, Query 814 messages), these are UDP encapsulated with an IP router alert option. 815 Each signaling node will therefore 'see' all such messages. The case 816 where the NSLPID does not match a local signaling application is 817 considered below in Section 4.3.4; otherwise, it is passed up to the 818 GIMPS layer for further processing as in the other cases. 820 4.3.2 Local Processing 822 Once a message has been received, by any method, it is processed 823 locally within the GIMPS layer. The GIMPS processing to be done 824 depends on the payloads carried; most of the GIMPS-internal payloads 825 are associated with state maintenance and are covered in Section 4.4. 827 One GIMPS-internal payload which is carried in each message and 828 requires processing is the GIMPS hop count. This is decremented on 829 input processing, and checked to be greater than zero on output 830 processing. The primary purpose of the GIMPS hop count is to prevent 831 message looping. 833 The remainder of the GIMPS message consists of an NSLP payload. This 834 is delivered locally to the signaling application identified at the 835 GIMPS level; the format of the NSLP payload is not constrained by 836 GIMPS, and the content is not interpreted. 838 Signaling applications can generate their messages for transmission, 839 either asynchronously, or in response to an input message, and GIMPS 840 can also generate messages autonomously. Regardless of the source, 841 outgoing messages are passed downwards for message transmission. 843 4.3.3 Message Transmission 845 When a message is available for transmission, GIMPS uses internal 846 policy and the stored routing state to determine how to handle it. 847 The following processing applies equally to locally generated 848 messages and messages forwarded from within the GIMPS or signaling 849 application levels. 851 The main decision is whether the message must be sent in connection 852 mode or datagram mode. Reasons for using the former could be: 854 o NSLP requirements: for example, the signaling application has 855 requested channel secured delivery, or reliable delivery; 857 o protocol specification: for example, this document specifies that 858 a message that requires fragmentation MUST be sent over a 859 messaging association; 861 o local GIMPS policy: for example, a node may prefer to send 862 messages over a messaging association to benefit from adaptive 863 congestion control. 865 In principle, as well as determining that some messaging association 866 must be used, GIMPS could select between a set of alternatives, e.g. 867 for load sharing or because different messaging associations provide 868 different transport or security attributes. 870 If the use of a messaging association is selected, the message is 871 queued on the association (found from the upstream or downstream peer 872 state table), and further output processing is carried out according 873 to the details of the protocol stack used for the association. If no 874 appropriate association exists, the message is queued while one is 875 created (see Section 4.4). If no association can be created, this is 876 an error condition, and should be indicated back to the NSLP. 878 If a messaging association is not required, the message is sent in 879 datagram mode. The processing in this case depends on the message 880 type and whether routing state exists or not. 882 o If the message is not a Query, and routing state exists, it is UDP 883 encapsulated and sent directly to that address. 885 o If the message is a Query, the it is UDP encapsulated with IP 886 address and router alert option determined from the MRI and NSLPID 887 (the details depend on the message routing method itself). 889 o If no routing state exists, GIMPS can attempt to use the same 890 IP/UDP encapsulation as in the Query case. If this is not 891 possible (e.g. because the encapsulation algorithm for the 892 message routing method is only defined valid for one message 893 direction), then this is an error condition which is reported back 894 to the local signaling application. 896 4.3.4 Bypass Forwarding 898 A GIMPS node may have to handle messages for which it has no 899 signaling application corresponding to the message NSLPID. There are 900 several possible cases depending mainly on the RAO setting (see 901 Section 5.3.2.1 for more details): 903 1. A datagram mode message contains an RAO value which is relevant 904 to NSIS but not the specific node, but the IP layer is unable to 905 recognise whether it needs to be passed to GIMPS for further 906 processing or whether the packet should be forwarded just like a 907 normal IP datagram. 909 2. A datagram mode message contains an RAO value which is relevant 910 to the node, but the specific signaling application for the 911 actual NSLPID in the message is not processed there. 913 3. A message is delivered directly to the node for which there is no 914 corresponding signaling application. (According to the rules of 915 the current specification, this should never happen. However, 916 future versions might find a use for such a feature.) 918 In all cases, the role of GIMPS is to forward the message essentially 919 unchanged. However, a GIMPS implementation must ensure that the IP 920 TTL field and GIMPS hop count are managed correctly to prevent 921 message looping, and this should be done consistently independently 922 of whether the processing (e.g. for case (1)) takes place on the 923 fast path or in GIMPS-specific code. The rules are that in cases (1) 924 and (2), the IP TTL is decremented just as if the message was a 925 normal IP forwarded packet; in cases (2) and (3) the GIMPS hop count 926 is decremented as in the case of normal input processing. These 927 rules are summarised in the following table: 929 +-------------+-------------+-------------------+-------------------+ 930 | Match RAO? | Match | IP TTL Handling | GHC Handling | 931 | | NSLPID? | | | 932 +-------------+-------------+-------------------+-------------------+ 933 | No | N/A (NSLPID | Decrement; | Ignore | 934 | | not | forward message | | 935 | | examined) | | | 936 | | | | | 937 | Yes | No | Decrement; | Decremented | 938 | | | forward message | | 939 | | | | | 940 | Message | No | Reset | Decrement and | 941 | directly | | | forward at GIMPS | 942 | addressed | | | level (not | 943 | | | | possible in | 944 | | | | current | 945 | | | | specification) | 946 | | | | | 947 | Yes, or | Yes | Locally delivered | N/A (ignored) | 948 | message | | | | 949 | directly | | | | 950 | addressed | | | | 951 +-------------+-------------+-------------------+-------------------+ 953 4.4 Routing State and Messaging Association Maintenance 955 The main responsibility of the GIMPS layer is to manage the routing 956 state and messaging associations which are used in the basic message 957 processing described above. Routing state is installed and 958 maintained by datagram mode messages containing specific GIMPS 959 payloads. Messaging associations are dependent on the existence of 960 routing state, but are actually set up by the normal procedures of 961 the transport and security protocols that comprise them. Timers 962 control routing state and messaging association refresh and 963 expiration. 965 There are two different cases for state installation and refresh: 967 1. Where routing state is being discovered or a new association is 968 to be established; and 970 2. Where an existing association can be re-used, including the case 971 where routing state for the association is being refreshed. 973 These cases are now considered in turn, along with the case of 974 general management procedures. 976 4.4.1 State Setup 978 The complete sequence of possible messages for state setup between 979 adjacent peers is shown in Figure 4 and described in detail in the 980 following text. 982 The initial message in any routing state maintenance operation is a 983 GIMPS-Query message, sent from the querying node and intercepted at 984 the responding node. This has addressing and other identifiers 985 appropriate for the flow and signaling application that state 986 maintenance is being done for, addressing information about the node 987 itself, and it is allowed to contain an NSLP payload. The querying 988 node also includes additional payloads: a Query Cookie, and 989 optionally a proposal for possible messaging association protocol 990 stacks. The role of the cookies in this and subsequent messages is 991 to protect against certain denial of service attacks and to correlate 992 the various events in the message sequence. 994 In the responding node, the GIMPS level processing of the GIMPS-Query 995 triggers the generation of a 'GIMPS-Response' message. This is a 996 'normally' encapsulated datagram mode message with additional 997 payloads. It contains addressing information about the responding 998 node, it echoes the Query Cookie, and can contain an NSLP payload 999 (possibly a response to the NSLP payload in the initial message). In 1000 case a messaging association was requested, it must also contain a 1001 Responder Cookie and counter proposal for the stack configuration. 1002 Otherwise, it may still include a Responder Cookie if the node's 1003 routing state setup policy requires it (see below). 1005 +----------+ +----------+ 1006 | Querying | |Responding| 1007 | Node | | Node | 1008 +----------+ +----------+ 1009 GIMPS-query 1010 ----------------------> ............. 1011 Router Alert Option . Routing . 1012 MRI/SID/NSLPID . state . 1013 Q-Node Addressing . installed . 1014 Query Cookie . at . 1015 [Q-Stack Proposal] . R-node(1) . 1016 [NSLP Payload] ............. 1018 ...................................... 1019 . The responder can use an existing . 1020 . messaging association if available . 1021 . from here onwards to short-circuit . 1022 . messaging association setup . 1023 ...................................... 1025 GIMPS-response 1026 ............. <---------------------- 1027 . Routing . MRI/SID/NSLPID 1028 . state . R-Node Addressing (D Mode only) 1029 . installed . Query cookie 1030 . at . [R-Stack Proposal] 1031 . Q-node . [Responder Cookie] 1032 ............. [NSLP Payload] 1034 .................................... 1035 . If a messaging association needs . 1036 . to be created, it is set up here . 1037 .................................... 1039 GIMPS-confirm 1040 ----------------------> 1041 MRI/SID/NSLPID 1042 Q-Node Addressing (D Mode only) 1043 Responder Cookie ............. 1044 [R-Stack Proposal] . Routing . 1045 [NSLP Payload] . state . 1046 . installed . 1047 . at . 1048 . R-node(2) . 1049 ............. 1051 Figure 4: Message Sequence at State Setup 1053 Setup of a new messaging association begins when both peer addressing 1054 information is available at the Querying node, and a new messaging 1055 association is actually needed. The setup has to be contemporaneous 1056 with a specific GIMPS-Query/Response exchange, because the addressing 1057 information used may have a limited lifetime (either because it 1058 depends on limited lifetime NAT bindings, or because it refers to 1059 agile destination ports for the transport protocols). Setup of the 1060 messaging association always starts from the Querying node, but the 1061 association itself can be used equally in both directions. 1063 The GIMPS-Confirm is the first message sent over the association and 1064 echoes the Responder Cookie and Stack Proposal from the 1065 GIMPS-Response (the latter is to prevent certain bidding-down attacks 1066 on messaging association security); the assocation can be used in the 1067 upstream direction for that flow and NSLPID after the Confirm has 1068 been received. The negotiation of what protocols to use for the 1069 messaging association is controlled by the Stack-Proposal and 1070 Node-Addressing information exchanged, and the processing of these 1071 objects is described in more detail in Section 5.5. 1073 The querying node installs the responder address as peer state 1074 information after verifying the Query Cookie in the GIMPS-Response. 1075 The responding node can install the querying address as peer state 1076 information at two points in time: 1078 1. after the receipt of the initial GIMPS-Query, or 1080 2. after a GIMPS-Confirm message containing the Responder Cookie. 1082 The detailed constraints on precisely when state information is 1083 installed are driven by local policy driven by security 1084 considerations on prevention of denial-of-service attacks and state 1085 poisoning attacks, which are discussed further in Section 7. 1087 4.4.2 Association Re-use 1089 It is a general design goal of GIMPS that, so far as possible, 1090 messaging associations should be re-used for multiple flows and 1091 sessions, rather than a new association set up for each. This is to 1092 ensure that the association cost scales like the number of peers 1093 rather than the number of flows or messages, and to avoid the latency 1094 of new association setup where possible. 1096 However, association re-use requires the identification of an 1097 existing association which matches the same routing state and desired 1098 properties that would be the result of a full handshake in D-mode, 1099 and this identification must be done as reliably and securely as 1100 continuing with the full procedure. Note that this requirement is 1101 complicated by the fact that NATs may remap the node addresses in 1102 D-mode messages, and also interacts with the fact that some nodes may 1103 peer over multiple interfaces (with different addresses). 1105 Association re-use is controlled by two fields in the Node-Addressing 1106 object (NAO), which is carried in GIMPS-query and GIMPS-response 1107 messages. The NAO includes: 1109 Peer-Identity: For a given node, this is a stable quantity (interface 1110 independent) with opaque syntax. It should be chosen so as to 1111 have a high probability of uniqueness between peers. Note that 1112 there is no cryptographic protection of this identity (attempting 1113 to provide this would essentially duplicate the functionality in 1114 the messaging association security protocols). 1116 Interface-Address: This is an IP address associated with the 1117 interface through which the flow associated with the signaling is 1118 routed. This can be considered as a routable identifier through 1119 which the signaling node can be reached; further discussion is 1120 contained in Section 5.5. 1122 By default, a messaging association is associated with the NAO that 1123 was provided by the peer at the time the assocation was set up. 1124 There may be more than one association for a given NAO (e.g. with 1125 different properties). 1127 Association re-use is controlled by matching the NAO provided in a 1128 GIMPS message with those associated with existing associations. This 1129 can be done on receiving either a GIMPS-Query or GIMPS-Response (the 1130 former is more likely): 1132 o If there is a perfect match to the NAO of an existing association, 1133 that association can be re-used (provided it has the appropriate 1134 properties in other respects). This is indicated by sending the 1135 following messages in the handshake over that association, 1136 omitting the NAO information. This will only fail (i.e. lead to 1137 re-use of an assocation to the 'wrong' node) if signaling nodes 1138 have colliding Peer-Identities, and one is reachable at the same 1139 Interface-Address as another. (This could be done by an on-path 1140 attacker.) 1142 o In all other cases, the full handshake is executed in datagram 1143 mode as usual. There are in fact four cases: 1145 1. Nothing matches: this is clearly a new peer. 1147 2. Only the Peer-Identity matches: this may be either a new 1148 interface on an existing peer, or a changed address mapping 1149 behind a NAT, or an attacker attempting to hijack the 1150 Peer-Identity. These should be rare events, so the expense of 1151 a new assocation setup is acceptable. If the authenticated 1152 peer identities match after assocation setup, the two 1153 Interface-Addresses may be bound to the assocation. 1155 3. Only the Interface-Address matches: this is probably a new 1156 peer behind the same NAT as an existing one. A new assocation 1157 setup is required. 1159 4. The full NAO matches: this is a degenerate case, where one 1160 node recognises an existing peer, but wishes to allow the 1161 option to set up a new association in any case. 1163 4.4.3 Background Maintenance 1165 Refresh and expiration of all types of state is controlled by timers. 1166 State in the routing table has a per-flow, per-direction timer, which 1167 expires after a routing state lifetime. It is the responsibility of 1168 the Querying node to generate a GIMPS-Query message before this timer 1169 expires, if it believes that the flow is still active. Receipt of 1170 the message at the responding node will refresh peer addressing state 1171 for one direction, and receipt of a GIMPS-Response at the querying 1172 node will refresh it for the other. Note that responding nodes do 1173 not control the refresh of routing state themselves, they are 1174 dependent on their peer for this. 1176 Messaging associations can be managed by either end; management 1177 consists of tearing down unneeded associations. Whether an 1178 association is needed is a local policy decision, which could take 1179 into account the cost of keeping the messaging association open, the 1180 level of past activity on the association, and the likelihood of 1181 future activity (e.g. if there are flows still in place which might 1182 generate messages that would use it). Messaging associations can 1183 always be set up on demand, and messaging association status is not 1184 made directly visible outside the GIMPS layer. Therefore, even if 1185 GIMPS tears down and later re-establishes a messaging association, 1186 signaling applications cannot distinguish this from the case where 1187 the association is kept permanently open. (To maintain the transport 1188 semantics decribed in Section 4.1, GIMPS must close transport 1189 connections carrying reliable messages gracefully or report an error 1190 condition, and must not open a new association for a given session 1191 and peer while messages on a previous association may still be 1192 outstanding.) 1194 5. Message Formats and Transport 1196 5.1 GIMPS Messages 1198 All GIMPS messages begin with a common header, which includes a 1199 version number, information about message type, signaling 1200 application, and additional control information. The remainder of 1201 the message is encoded in an RSVP-style format, i.e., as a sequence 1202 of type-length-value (TLV) objects. This subsection describes the 1203 possible GIMPS messages and their contents at a high level; a more 1204 detailed description of each information element is given in 1205 Section 5.2. 1207 The following gives the syntax of GIMPS messages in ABNF [3]. 1209 GIMPS-Message: A message is either a one of the stages in the 3-way 1210 handshake, or a simple message carrying NSLP data. 1212 GIMPS-Message = GIMPS-Query / GIMPS-Response / 1213 GIMPS-Confirm / GIMPS-Data 1215 GIMPS-Query: A GIMPS-Query is always sent in datagram mode. As well 1216 as the common header, it contains certain mandatory control objects, 1217 and may contain a signaling application payload. A stack proposal is 1218 mandatory if the message exchange relates to setup of a messaging 1219 association. 1221 GIMPS-Query = Common-Header 1222 Message-Routing-Information 1223 Session-Identification 1224 Node-Addressing 1225 Query-Cookie 1226 [ Stack-Proposal ] 1227 [ Routing-State-Lifetime ] 1228 [ NSLP-Data ] 1230 GIMPS-Response: A GIMPS-Response may be sent in datagram or 1231 connection mode (if a messaging association is being re-used). It 1232 echoes the MRI, SID and Query-Cookie of the Query, and carries its 1233 own Node-Addresing information; if the message exchange relates to 1234 setup of a messaging association (which can only take place in 1235 datagram mode), a Responder cookie is mandatory, and it must also 1236 contain its own Stack-Proposal. 1238 GIMPS-Response = Common-Header 1239 Message-Routing-Information 1240 Session-Identification 1241 Node-Addressing 1242 Query-Cookie 1243 [ Responder-Cookie [ Stack-Proposal ] ] 1244 [ Routing-State-Lifetime ] 1245 [ NSLP-Data ] 1247 GIMPS-Confirm: A GIMPS-Confirm may be sent in datagram or connection 1248 mode (if a messaging association has been re-used). It echoes the 1249 MRI, SID and Responder-Cookie of the Response; if the message 1250 exchange relates to setup of a new messaging association or reuse of 1251 an existing one (which can only take place in connection mode), the 1252 message must also echo the Stack-Proposal from the GIMPS-Response so 1253 it can be verified that this has not been tampered with. 1255 GIMPS-Confirm = Common-Header 1256 Message-Routing-Information 1257 Session-Identification 1258 Node-Addressing 1259 Responder-Cookie 1260 [ Stack-Proposal ] 1261 [ Routing-State-Lifetime ] 1262 [ NSLP-Data ] 1264 GIMPS-Data: A plain data message contains no control objects, but 1265 only the MRI and SID assocated with the NSLP data being transferred. 1266 Node-Addressing information is only carried in the datagram mode 1267 case. 1269 GIMPS-Data = Common-Header 1270 Message-Routing-Information 1271 Session-Identification 1272 [ Node-Addressing ] 1273 NSLP-Data 1275 5.2 Information Elements 1277 This section describes the content of the various information 1278 elements that can be present in each GIMPS message, both the common 1279 header, and the individual TLVs. The format description in terms of 1280 bit patterns is provided in Appendix C. 1282 5.2.1 The Common Header 1284 Each message begins with a fixed format common header, which contains 1285 the following information: 1287 Version: The version number of the GIMPS protocol. 1289 Length: The number of words in the message following the common 1290 header. 1292 Signaling application identifier (NSLPID): This describes the 1293 specific signaling application, such as resource reservation or 1294 firewall control. 1296 GIMPS hop counter: A hop counter to prevent a message from looping 1297 indefinitely. 1299 Message type: The message type (Query, Response, etc.) 1301 Source addressing mode: A flag to indicate whether the IP source 1302 address of the message was set to be the signaling source address, 1303 or whether it was derived from the message routing information in 1304 the payload. 1306 5.2.2 TLV Objects 1308 All data following the common header is encoded as a sequence of 1309 type-length-value objects. Currently, each object can occur at most 1310 once; the set of required and permitted objects is determined by the 1311 message type and further information in the common header. 1313 These items are contained in each GIMPS message: 1315 Message-Routing-Information (MRI): Information sufficient to define 1316 how the signaling message should be routed through the network. 1318 Message-Routing-Information = message-routing-method 1319 method-specific-information 1321 The format of the method-specific-information depends on the 1322 message-routing-method requested by the signaling application. In 1323 the basic path-coupled case, it is just the Flow Identifier as in 1324 [20]. Minimally, this could just be the flow destination address; 1325 however, to account for policy based forwarding and other issues a 1326 more complete set of header fields should be used (see Section 6.2 1327 and Section 6.3 for further discussion). 1329 The MRI is essentially a read only object for GIMPS processing. 1330 It is set by the NSLP in the message sender and used by GIMPS to 1331 select the message addressing, but not otherwise modified. Note 1332 that every message routing method must implicitly define a 1333 directionality (upstream vs. downstream), corresponding to the 1334 two directions in the routing state table, and the MRI must 1335 include control information which says in which direction this 1336 message is being sent. 1338 Flow-Identifier = network-layer-version 1339 source-address prefix-length 1340 destination-address prefix-length 1341 IP-protocol 1342 traffic-class 1343 [ flow-label ] 1344 [ ipsec-SPI / L4-ports] 1346 Additional control information defines whether the flow-label, SPI 1347 and port information are present, the direction of the message 1348 relative to this flow, and whether the IP-protocol and 1349 traffic-class fields should be interpreted as significant. 1351 Session-Identification (SID): The GIMPS session identifier is a long, 1352 cryptographically random identifier chosen by the node which 1353 originates the signaling exchange. The length is open, but 128 1354 bits should be more than sufficient to make the probability of 1355 collisions orders of magnitude lower than other failure reasons. 1356 The session identifier should be considered immutable end-to-end 1357 along the flow path (GIMPS never changes it, and signaling 1358 applications should propagate it unchanged on messages for the 1359 same session). 1361 The following items are optional: 1363 Node addressing: This can include a peer identity and IP address for 1364 the sending node, as well as higher layer addressing information 1365 for the negotiation of messaging association protocols. It also 1366 includes IP TTL information to allow the hop count between GIMPS 1367 peers to be measured and reported. 1369 Node-Addressing = peer-identity 1370 IP-TTL 1371 [ interface-address ] 1372 [ *higher-layer-addressing ] 1374 The peer-identity and interface-address are used for matching 1375 existing associations, as discussed in Section 4.4.2. Any 1376 technique may be used to generate it, so long as it is stable. 1377 The interface-address should be a routable address where the 1378 sending node can be reached over UDP or messaging association 1379 protocols. Where this object is used in a GIMPS-Query, it should 1380 specifically be set to the address of the interface that will be 1381 used for the outbound flow, to allow its use in route change 1382 handling, see Section 6.1. The purpose and structure of the 1383 higher-layer-addressing fields is described in Section 5.5. Note 1384 that the higher-layer-addressing fields are only present in 1385 datagram encapsulated messages; when this object is carried in 1386 connection mode, these information elements are neither necessary 1387 or meaningful. 1389 The setting and interpretation of the IP-TTL field depends on the 1390 message direction (as determined from the MRI) and encapsulation. 1392 * If the message is downstream, the IP-TTL is set to the TTL that 1393 will be set in the IP header for the message (if this can be 1394 determined), or else 0. 1396 * On receiving a downstream message in datagram mode, the IP-TTL 1397 is compared to the TTL in the IP header, and the result is 1398 stored as the IP-hop-count-to-peer for the upstream peer in the 1399 routing state table for that flow. Otherwise, the field is 1400 ignored. 1402 * If the message is upstream, the IP-TTL is set to the value of 1403 the IP-hop-count-to-peer stored in the routing state table, or 1404 0 if there is no value yet stored. 1406 * On receiving an upstream message, the IP-TTL is stored as the 1407 IP-hop-count-to-peer for the downstream peer. 1409 In all cases, the TTL value reported to signaling applications is 1410 the one stored with the routing state for that flow, after it has 1411 been updated (if appropriate) from processing the message in 1412 question. 1414 Stack Proposal: This field contains information about which 1415 combinations of transport and security protocols are proposed for 1416 use in messaging associations, and is also discussed further in 1417 Section 5.5. 1419 Stack-Proposal = *stack-profile 1421 stack-profile = *protocol-layer 1423 Each protocol-layer field identifies a protocol with a unique 1424 tag; any address-related (mutable) information associated with the 1425 protocol will be carried in a higher-layer-addressing field in the 1426 Node-Addressing TLV (see above). 1428 Query-Cookie/Responder-Cookie: A query-cookie is contained in a 1429 GIMPS-Query message and must be echoed in a GIMPS-Response; a 1430 response-cookie is optional in a GIMPS-Response message, and if 1431 present must be echoed in the following GIMPS-Confirm message. 1432 Cookies are variable length (chosen by the cookie generator) and 1433 need to be designed so that a node can determine the validity of a 1434 cookie without keeping state. A future version of this 1435 specification will include references to techniques for generating 1436 such cookies. 1438 Routing-State-Lifetime: The lifetime of GIMPS routing state in the 1439 absence of refreshes, measured in seconds. Defaults to 30 1440 seconds. 1442 NSLP-Data: The NSLP payload to be delivered to the signaling 1443 application. GIMPS does not interpret the payload content. 1445 5.3 Datagram Mode Transport 1447 This section describes the various encapsulation options for datagram 1448 mode messages. Although there are several variant possibilities, 1449 depending on message type, message routing method, and local policy, 1450 the general design principle is that the sole purpose of the 1451 encapsulation is to ensure that the message is delivered to or 1452 intercepted at the correct peer. Beyond that, no significance is 1453 attached to the type of encapsulation or the values of addresses or 1454 ports used for it. This allows new options to be developed in the 1455 future to handle particular deployment requirements without modifying 1456 the overall protocol specification. 1458 5.3.1 Normal Encapsulation 1460 Normal encapsulation is used for all datagram mode messages where the 1461 signaling peer is already known from previous signaling. This 1462 includes Response and Confirm messages, and Data messages except if 1463 these are being sent without using local routing state. Normal 1464 encapsulation is simple: the complete set of GIMPS payloads is 1465 concatenated together with the common header, and placed in the data 1466 field of a UDP datagram. UDP checksums should be enabled. The 1467 message is IP addressed directly to the adjacent peer; the UDP port 1468 numbering should be compatible with that used on Query messages (see 1469 below), that is, the same for messages in the same direction and 1470 swapped otherwise. 1472 5.3.2 Query Encapsulation 1474 Query encapsulation is used for messages where no routing state is 1475 available or where the routing state is being refreshed, in 1476 particular for GIMPS-Query messages. Query encapsulation is similar 1477 to normal encapsulation, with changes in IP address selection, IP 1478 options, and a defined method for selecting UDP ports. 1480 In general, the IP addresses are derived from information in the MRI; 1481 the exact rules depend on the message routing method. In addition, 1482 the IP header is given a Router Alert Option to assist the peer in 1483 intercepting the message depending on the NSLPID. Router alert 1484 option value-field setting is discussed in Section 5.3.2.1. 1486 The source UDP port is selected by the message sender as the port at 1487 which it is prepared to receive UDP messages in reply, and a 1488 destination UDP port should be allocated by IANA. Note that GIMPS 1489 may send messages addressed as {flow sender, flow receiver} which 1490 could make their way to the flow receiver even if that receiver were 1491 GIMPS-unaware. This should be rejected (with an ICMP message) rather 1492 than delivered to the user application (which would be unable to use 1493 the source address to identify it as not being part of the normal 1494 data flow). Therefore, a "well-known" port would seem to be 1495 required. 1497 5.3.2.1 Intermediate Node Bypass and Router Alert Values 1499 We assume that the primary mechanism for intercepting messages is the 1500 use of the RAO. The RAO contains a 16 bit value field, within which 1501 35 values have currently been assigned by IANA. This section 1502 discusses the technical considerations to be taken into account when 1503 assigning values for use by GIMPS. 1505 The basic goal is to optimise protocol processing, i.e. to minimise 1506 the amount of slow-path processing that nodes have to carry out for 1507 messages they are not actually interested in. There are two basic 1508 reasons why a GIMPS node might wish to ignore a message: 1510 o because it is for a signaling application that the node does not 1511 process; 1513 o because even though the signaling application is present on the 1514 node, the interface on which the message arrives is only 1515 processing signaling messages at the aggregate level and not for 1516 individual flows (compare [15]). 1518 Conversely, note that a node might wish to process a number of 1519 different signaling applications, either because it was genuinely 1520 multifunctional or because it processed several versions of the same 1521 application. (Note from Appendix C.1 that different versions are 1522 distinguished by different NSLP identifiers.) 1523 Some or all of this information can be encoded in the RAO value 1524 field, which then allows messages to be filtered on the fast path. 1525 There is a tradeoff between two approaches here, whose evaluation 1526 depends on whether the processing node is specialised or general 1527 purpose: 1529 Fine-Grained: The signaling application (including specific version) 1530 and aggregation level are directly identified in the RAO value. A 1531 specialised node which handles only a single NSLP can efficiently 1532 ignore all other messages; a general purpose node may have to 1533 match the RAO value in a message against a long list of possible 1534 values. 1536 Coarse-Grained: IANA allocates RAO values for 'popular' applications 1537 or groups of applications (such as 'All QoS Signaling 1538 Applications'). This speeds up the processing in a general 1539 purpose node, but a specialised node may have to carry out further 1540 processing on the GIMPS common header to identify the precise 1541 messages it needs to consider. 1543 These considerations imply that the RAO value should not be tied 1544 directly to the NSLP id, but should be selected for the application 1545 on broader considerations of likely deployment scenarios. Note that 1546 the exact NSLP is given in the GIMPS common header, and some 1547 implementations may still be able to process it on the fast path. 1548 The semantics of the node dropping out of the signaling path are the 1549 same however the filtering is done (see Section 4.3.4). 1551 There is a special consideration in the case of the aggregation 1552 level. In this case, whether a message should be processed depends 1553 on the network region it is in (specifically, the link it is on). 1554 There are then two basic possibilities: 1556 1. All routers have essentially the same algorithm for which 1557 messages they process, i.e. all messages at aggregation level 0. 1558 However, messages have their aggregation level incremented on 1559 entry to an aggregation region and decremented on exit. 1561 2. Router interfaces are configured to process messages only above a 1562 certain aggregation level and ignore all others. The aggregation 1563 level of a message is never changed; signaling messages for end 1564 to end flows have level 0, but signaling messages for aggregates 1565 are generated with a higher level. 1567 The first technique requires aggregating/deaggregating routers to be 1568 configured with which of their interfaces lie at which aggregation 1569 level, and also requires consistent message rewriting at these 1570 boundaries. The second technique eliminates the rewriting, but 1571 requires interior routers to be configured also. It is not clear 1572 what the right trade-off between these options is. 1574 5.3.2.2 Query Encapsulation for the Path-Coupled Message Routing Method 1576 For the case of the path-coupled message routing method, where the 1577 message is travelling in the same ('downstream') direction as the 1578 flow defined by the MRI, the IP addressing for Query messages is as 1579 follows: 1581 o The destination address MUST be the flow destination address as 1582 given in the MRI of the message payload. 1584 o By default, the source address is the flow source address, again 1585 from the message MRI. This provides the best likelihood that the 1586 message will be correctly routed through any region which performs 1587 per-packet policy-based forwarding or load balancing which takes 1588 the source address into account. However, there may be 1589 circumstances where the use of the signaling source address is 1590 preferable, specifically: 1592 * In order to receive ICMP error messages about the Query message 1593 (specifically, unreachable port or address). If these are 1594 delivered to the flow source rather than the signaling source, 1595 it will be very difficult for the querying node to detect that 1596 it is the last GIMPS node on the path. 1598 * In order to attempt to run GIMPS through an unmodified NAT, 1599 which will only process and translate IP addresses in the IP 1600 header. 1602 Because of these considerations, use of the signaling source 1603 address is allowed as an option, which is use based on local 1604 policy. A node SHOULD use the flow source address for initial 1605 Query messages, but MAY transition to the signaling source address 1606 for retransmissions or as a matter of static configuration (e.g. 1607 if a NAT is known to be in the path out of a certain interface). 1608 A flag in the common header tells the message receiver which 1609 option was used. 1611 It is vital that the Query message truly mimics the actual data flow, 1612 since this is the basis of how the signaling message is attached to 1613 the data path. To this end, GIMPS may set the traffic class and (for 1614 IPv6) flow label to match the values in the Flow-Identifier if this 1615 would be needed to ensure correct routing. 1617 These encapsulation rules allow Query messages to be sent in the same 1618 direction as the flow, and hence allow routing state to be set up 1619 from the flow source towards the flow destination. In some 1620 deployment scenarios (see Section 9.1 for further discussion), it is 1621 desirable and logically possible to set up routing state in the 1622 reverse direction. Implementing this in the specification would 1623 require defining rules for encapsulating a Query message in the 1624 upstream direction. Details are for further study. 1626 5.3.3 Retransmission and Rate-Control 1628 Datagram mode is built on top of UDP, and hence has no automatic 1629 reliability or congestion control capabilities. Signaling 1630 applications requiring reliability should be serviced using C-mode, 1631 which should also carry the bulk of signaling traffic. However, some 1632 form of messaging reliability is required for the GIMPS control 1633 messages themselves, as is rate control to handle retransmissions and 1634 also bursts of unreliable signaling or state setup requests from the 1635 signaling applications. 1637 GIMPS-Query messages which do not receive GIMPS-responses should be 1638 retransmitted with a binary exponential backoff, with an initial 1639 timeout of T1 up to a maximum of T2 seconds. The values of T1 and T2 1640 may be implementation defined; default values are for further study. 1641 The value of T1 may be increased on long latency links. Note that 1642 GIMPS-Queries may go unanswered either because of message loss, or 1643 because there is no reachable GIMPS peer. Therefore, implementations 1644 must trade off reliability (large T2) against promptness of error 1645 feedback to applications (small T2). GIMPS-Responses should always 1646 be sent promptly to avoid spurious retransmissions. Retransmitted 1647 GIMPS-Queries should use different Query-Cookie values and will 1648 therefore elicit different GIMPS-Responses. If either message 1649 carries NSLP data, it may be delivered multiple times to the 1650 signaling application. 1652 Other datagram mode messages are not generally retransmitted. 1653 GIMPS-Responses do not need reliability; if they are lost, the 1654 initiating Query will eventually be resent. 1656 The case of a lost GIMPS-Confirm is more subtle. Notionally, we can 1657 distinguish between two cases: 1659 o Where the Responding node is already prepared to store per-flow 1660 state after receiving a single (Query) message. This would 1661 include any cases where the node has NSLP data queued to send. 1662 Here, it is reasonable for the protocol to demand that the 1663 Responding node runs a retransmission timer to resend the Response 1664 message until a Confirm is received. The problem of an 1665 amplification attack stimulated by a malicious Query should be 1666 handled by requiring the cookie mechanism to enable the node 1667 receiving the Response to discard it efficiently if it does not 1668 match a previously sent Query. 1670 o where the responding node is not prepared to store per-flow state 1671 until receiving a properly formed Confirm message. 1673 The second (which is probably the more commonplace one where Confirm 1674 messages are wanted at all), a retransmission timer should not be 1675 required. However, we can assume that the next signaling message 1676 will be in the direction Querying Node -> Responding Node (if there 1677 is no 'next signaling message' the fact that the Confirm has been 1678 lost is moot). In this case, the responding node will start to 1679 receive messages at the GIMPS level for a flow/NSLP combination for 1680 which there is no stored routing state (since this state is only 1681 created on receipt of a Confirm). 1683 The consequence of this is that the error condition is detected at 1684 the Responding node when such a message arrives without the need for 1685 a specific timer. Recovery requires a Confirm to be retransmitted 1686 and successfully received. The ideal mechanism to cause this would 1687 be for the Responding node to be able to reject the incoming message 1688 with an error "No Routing State Exists" back to the Querying node, 1689 which would interpret this as caused by a lost Confirm; the Querying 1690 node needs to be able to regenerate the Confirm from local state 1691 without getting a Response (e.g. in particular it needs to remember 1692 the Responder Cookie value). 1694 The basic rate limiting requirements for datagram mode traffic are 1695 deliberately minimal. A single rate limiter applies to all traffic 1696 (for all interfaces and message types). It applies to 1697 retransmissions as well as new messages, although an implementation 1698 may choose to prioritise one over the other. When the rate limiter 1699 is imposed, datagram mode messages are queued until transmission is 1700 re-enabled, or an error condition may be indicated back to local 1701 signaling applications. The rate limiting mechanism is 1702 implementation defined, but it is recommended that a token bucket 1703 limiter as described in [8] should be used. 1705 5.4 Connection Mode Transport 1707 Encapsulation in connection mode is more complex, because of the 1708 variation in available transport functionality. This issue is 1709 treated in Section 5.4.1. The actual encapsulation is given in 1710 Section 5.4.2. 1712 5.4.1 Choice of Transport Protocol 1714 It is a general requirement of the NTLP defined in [20] that it 1715 should be able to support bundling (of small messages), fragmentation 1716 (of large messages), and message boundary delineation. Not all 1717 transport protocols natively support all these features. 1719 SCTP [6] satisfies all requirements. 1721 DCCP [7] is message based but does not provide bundling or 1722 fragmentation. Bundling can be carried out by the GIMPS layer 1723 sending multiple messages in a single datagram; because the common 1724 header includes length information (number of TLVs), the message 1725 boundaries within the datagram can be discovered during parsing. 1726 Fragmentation of GIMPS messages over multiple datagrams should be 1727 avoided, because of amplification of message loss rates that this 1728 would cause. 1730 TCP provides both bundling and fragmentation, but not message 1731 boundaries. However, the length information in the common header 1732 allows the message boundary to be discovered during parsing. 1734 The bundling together of small messages is either built into the 1735 transport protocol or can be carried out by the GIMPS layer during 1736 message construction. Either way, two approaches can be 1737 distinguished: 1739 1. As messages arrive for transmission they are gathered into a 1740 bundle until a size limit is reached or a timeout expires (cf. 1741 the Nagle algorithm of TCP or similar optional functionality in 1742 SCTP). This provides maximal efficiency at the cost of some 1743 latency. 1745 2. Messages awaiting transmission are gathered together while the 1746 node is not allowed to send them (e.g. because it is congestion 1747 controlled). 1749 The second type of bundling is always appropriate. For GIMPS, the 1750 first type is inappropriate for 'trigger' (i.e. state-changing) 1751 messages, but may be appropriate for refresh messages. These 1752 distinctions are known only to the signaling applications, but could 1753 be indicated (as an implementation issue) by setting the priority 1754 transfer attribute. 1756 It can be seen that all of these protocol options can be supported by 1757 the basic GIMPS message format already presented. GIMPS messages 1758 requiring fragmentation must be carried using a reliable transport 1759 protocol, TCP or SCTP. This specification defines only the use of 1760 TCP, but it can be seen that the other possibilities could be 1761 included without additional work on message formatting. 1763 5.4.2 Encapsulation Format 1765 The GIMPS message, consisting of common header and TLVs, is carried 1766 directly in the transport protocol (possibly incorporating transport 1767 layer security protection). Further GIMPS messages can be carried in 1768 a continuous stream (for TCP), or up to the next transport layer 1769 message boundary (for SCTP/DCCP/UDP). This situation is shown in 1770 Figure 5; it applies to both upstream and downstream messages. 1772 +---------------------------------------------+ 1773 | L2 Header | 1774 +---------------------------------------------+ 1775 | IP Header | ^ 1776 | Source address = signaling source | ^ 1777 | Destination address = signaling destination | . 1778 +---------------------------------------------+ . 1779 | L4 Header | . ^ 1780 | (Standard TCP/SCTP/DCCP/UDP header) | . ^ 1781 +---------------------------------------------+ . . 1782 | GIMPS Message | . . ^ 1783 | (Common header and TLVs as in section 5.1) | . . ^ Scope of 1784 +---------------------------------------------+ . . . security 1785 | Additional GIMPS messages, each with its | . . . protection 1786 | own common header, either as a continuous | . . . (depending 1787 | stream, or continuing to the next L4 | . . . on channel 1788 . message boundary . . . . security 1789 . . V V V mechanism 1790 . . V V V in use) 1792 Figure 5: Connection Mode Encapsulation 1794 5.5 Messaging Association Negotiation 1796 5.5.1 Overview 1798 A key attribute of GIMPS is that it is flexible in its ability to use 1799 existing transport and security protocols. Different transport 1800 protocols may have performance attributes appropriate to different 1801 environments; different security protocols may fit appropriately with 1802 different authentication infrastructures. Even given an initial 1803 default mandatory protocol set for GIMPS, the need to support new 1804 protocols in the future cannot be ruled out, and secure protocol 1805 negotation cannot be added to an existing protocol in a 1806 backwards-compatible way. Therefore, some sort of protocol 1807 negotiation capability is required. 1809 Protocol negotiation is carried out in GIMPS-Query/Response messages, 1810 using Stack-Proposal and Node-Addressing objects. If a new messaging 1811 association is required it is then set up, followed by a 1812 GIMPS-Confirm. Messaging association re-use is achieved by 1813 short-circuiting this exchange by sending the GIMPS-Response or 1814 GIMPS-Confirm messages on an existing association (Section 4.4.2); 1815 whether to do this is a matter of local policy at the querying or 1816 responding node. It is always possible for a node to restrict itself 1817 to a single messaging association between two peers. If multiple 1818 associations exist, it is a matter of local policy how to distribute 1819 messages over them, subject to respecting the transfer attributes 1820 requested. 1822 The end result of the negotiation is a messaging assocation which is 1823 a stack of protocols. Every possible protocol has the following 1824 attributes: 1826 o A Protocol-Identifier, a 1-byte IANA assigned value. 1828 o A specification of the (non-negotiable) policies about how the 1829 protocol should be used (for example, connection open direction). 1831 o Formats for carrying the protocol addressing and other 1832 configuration information in higher-layer-addressing information 1833 elements. There are different formats depending on whether the 1834 information is carried in the Query or Response (the object for a 1835 Confirm echoes the Response). 1837 A Stack-Proposal object is simply a list of profiles; each profile is 1838 a sequence of Protocol-Identifiers. Stack-Proposals are generally 1839 accompanied by Node-Addressing objects; as well as a Peer-Identity 1840 and Interface-Address, this carries a higher-layer-addressing 1841 information element for every protocol listed in the Stack-Proposal. 1842 A node generating a Node-Addressing object is committed to honouring 1843 the implied protocol configuration; in particular, it must be 1844 prepared to accept incoming datagrams or connections at the 1845 Interface-Address/protocol/port combinations advertised. However, 1846 the object contents should be retained only for the duration of the 1847 Query/Response exchange and any following association setup and 1848 afterwards discarded. (They may become invalid because of expired 1849 bindings at intermediate NATs, or because the advertising node is 1850 using agile ports.) 1852 A GIMPS-Query requesting association setup always contains a 1853 Stack-Proposal and Node-Addressing object, and unless re-use occurs, 1854 the GIMPS-Response does so also. For a GIMPS-Response, the 1855 Stack-Proposal must be invariant for the combination of outgoing 1856 interface and NSLPID (it must not depend on the GIMPS-Query). Once 1857 the messaging association is set up, the querying node repeats the 1858 responder's Stack-Proposal over it in the GIMPS-confirm. The 1859 resonding node can verify this to ensure that no bidding-down attack 1860 has occurred. Where the Response or Confirm is being sent in 1861 connection mode (either because of re-use or because messaging 1862 association setup has actually completed), the Node-Addressing object 1863 is sent in an abbreviated form, omitting the higher layer information 1864 fields. The Interface-Address is retained in the Confirm, to allow 1865 matching the messaging association against subsequent Query messages. 1867 5.5.2 Protocol Definition: Forwards-TCP 1869 This defines a basic configuration for the use of TCP between peers. 1870 Support for this protocol is mandatory; associations using it can 1871 carry messages with the transfer attribute Reliable=True. The 1872 connection is opened in the forwards direction, from the querying 1873 node, towards the responder at a previously advertised port. The 1874 higher-layer-addressing formats are: 1876 o downstream: no additional data (just the Protocol-Identifier) 1878 o upstream: 2 byte port number at which the connection will be 1879 accepted. 1881 5.5.3 Additional Protocol Options 1883 It is expected that the base GIMPS specification will define a single 1884 mandatory protocol for channel security (one of IKE/IPsec or TLS). 1885 Further protocols or configurations could be defined in the future 1886 for additional performance or flexibility. Examples are: 1888 o SCTP or DCCP as alternatives to TCP, with essentially the same 1889 configuration. 1891 o SigComp [17] for message compression. 1893 o ssh [25] or HIP/IPsec [26] for channel security. 1895 o Alternative modes of TCP operation, for example where it is set up 1896 from the responder to the querying node. 1898 6. Advanced Protocol Features 1900 6.1 Route Changes and Local Repair 1902 6.1.1 Introduction 1904 When re-routing takes place in the network, GIMPS and signaling 1905 application state needs to be updated for all flows whose paths have 1906 changed. The updates to signaling application state are usually 1907 signaling application dependent: for example, if the path 1908 characteristics have actually changed, simply moving state from the 1909 old to the new path is not sufficient. Therefore, GIMPS cannot carry 1910 out the complete path update processing. Its responsibilities are to 1911 detect the route change, update its own routing state consistently, 1912 and inform interested signaling applications at affected nodes. 1914 Route change management is complicated by the distributed nature of 1915 the problem. Consider the re-routing event shown in Figure 6. An 1916 external observer can tell that the main responsibility for 1917 controlling the updates will probably lie with nodes A and E; 1918 however, D1 is best placed to detect the event quickly at the GIMPS 1919 level, and B1 and C1 could also attempt to initiate the repair. 1921 On the assumption that NSLPs are soft-state based and operate end to 1922 end, and because GIMPS also periodically updates its picture of 1923 routing state, route changes will eventually be repaired 1924 automatically. However, especially if NSLP refresh times are 1925 extended to reduce signaling load, the duration of inconsistent state 1926 may be very long indeed. Therefore, GIMPS includes logic to deliver 1927 prompt notifications to NSLPs, to allow NSLPs to carry out local 1928 repair if possible. 1930 xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx 1931 x +--+ +--+ +--+ x Initial 1932 x .|B1|_.......|C1|_.......|D1| x Configuration 1933 x . +--+. .+--+. .+--+\. x 1934 x . . . . . . x 1935 >>xxxxxx . . . . . . xxxxxx>> 1936 +-+ . .. .. . +-+ 1937 .....|A|/ .. .. .|E|_.... 1938 +-+ . . . . . . +-+ 1939 . . . . . . 1940 . . . . . . 1941 . +--+ +--+ +--+ . 1942 .|B2|_.......|C2|_.......|D2|/ 1943 +--+ +--+ +--+ 1945 +--+ +--+ +--+ Configuration 1946 .|B1|........|C1|........|D1| after failure 1947 . +--+ .+--+ +--+ of D1-E link 1948 . \. . \. ./ 1949 . . . . . 1950 +-+ . .. .. +-+ 1951 .....|A|. .. .. .|E|_.... 1952 +-+\. . . . . . +-+ 1953 >>xxxxxx . . . . . . xxxxxx>> 1954 x . . . . . . x 1955 x . +--+ +--+ +--+ . x 1956 x .|B2|_.......|C2|_.......|D2|/ x 1957 x +--+ +--+ +--+ x 1958 xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx 1960 ........... = physical link topology 1962 >>xxxxxxx>> = flow direction 1964 _.......... = indicates outgoing link 1965 for flow xxxxxx given 1966 by local forwarding table 1968 Figure 6: A Re-Routing Event 1970 6.1.2 Route Change Detection 1972 There are two aspects to detecting a route change at a single node: 1974 o Detecting that the path in the direction of the Query has (or may 1975 have) changed. 1977 o Detecting that the path in the direction of the Response has (or 1978 may have) changed (in which case the node may no longer be on the 1979 path at all). 1981 At a single node, these processes are largely independent, although 1982 clearly a change in the path in one direction at a node corresponds 1983 to a change in path in the opposite direction at its peer. Note that 1984 there are two possible aspects of route change: 1986 Interface: The interface through which a flow leaves or enters a node 1987 may change. 1989 Peer: The adjacent peer may change. 1991 In general, a route change could include one or the other or both. 1992 (In theory it could include neither, although such changes are hard 1993 to detect and even harder to do anything useful about.) 1995 There are five mechanisms for a GIMPS node to detect that a route 1996 change has occurred, which are listed below. They apply differently 1997 depending on whether the change is in the Query or Response 1998 direction, and these differences are summarised in the following 1999 table. 2001 Local Trigger: In trigger mode, a node finds out that the next hop 2002 has changed. This is the RSVP trigger mechanism where some form 2003 of notification mechanism from the routing table to the protocol 2004 handler is assumed. Clearly this only works if the routing change 2005 is local, not if the routing change happens somewhere a few 2006 routing hops away (including the case that the change happens at a 2007 GIMPS-unaware node). 2009 Extended Trigger: An extended trigger, where the node checks a 2010 link-state routing table to discover that the path has changed. 2011 This makes certain assumptions on consistency of route computation 2012 (but you probably need to make those to avoid routing loops) and 2013 only works within a single area for OSPF and similar link-state 2014 protocols. Where available, this offers the most accurate and 2015 expeditious indication of route changes, but requires more access 2016 to the routing internals than a typical OS may provide. 2018 GIMPS C-mode Monitoring: A node may find that C-mode packets are 2019 arriving (from either peer) with a different TTL or on a different 2020 interface. This provides no direct information about the new flow 2021 path, but indicates that routing has changed and that rediscovery 2022 may be required. 2024 Data Plane Monitoring: The signaling application on a node may detect 2025 a change in behaviour of the flow, such as TTL change, arrival on 2026 a different interface, or loss of the flow altogether. The 2027 signaling application on the node is allowed to notify this 2028 information locally to GIMPS. 2030 GIMPS Probing: In probing mode, each GIMPS node periodically repeats 2031 the discovery (GIMPS-Query/GIMPS-Response) operation. The 2032 querying node will discover the route change by a modification in 2033 the Node-Addressing information in the GIMPS-Response. This is 2034 similar to RSVP behavior, except that there is an extra degree of 2035 freedom since not every message needs to repeat the discovery, 2036 depending on the likely stability of routes. All indications are 2037 that, leaving mobility aside, routes are stable for hours and 2038 days, so this may not be necessary on a 30-second interval, 2039 especially if the other techniques listed above are available. 2041 When these methods discover a route change in the Response direction, 2042 this cannot be handled directly by GIMPS at the detecting node, since 2043 route discovery proceeds only in the Query direction. Therefore, to 2044 exploit these mechanisms, it must be possible for GIMPS to send a 2045 notification message to initiate this. (This would be possible for 2046 example by setting an additional flag in the Common-Header of a 2047 message.) 2049 +----------------------+----------------------+---------------------+ 2050 | Method | Query direction | Response direction | 2051 +----------------------+----------------------+---------------------+ 2052 | Local Trigger | Discovers new | Not applicable | 2053 | | interface (and peer | | 2054 | | if local) | | 2055 | | | | 2056 | Extended Trigger | Discovers new | May determine that | 2057 | | interface and may | route from peer | 2058 | | determine new peer | will have changed | 2059 | | | | 2060 | C-Mode Monitoring | Provides hint that | Provides hint that | 2061 | | change has occurred | change has occurred | 2062 | | | | 2063 | Data Plane | Not applicable | NSLP informs GIMPS | 2064 | Monitoring | | that a change may | 2065 | | | have occurred | 2066 | | | | 2067 | Probing | Discovers changed | Discovers changed | 2068 | | Node-Addressing in | Node-Addressing in | 2069 | | GIMPS-Response | GIMPS-Query | 2070 +----------------------+----------------------+---------------------+ 2072 6.1.3 Local Repair 2074 Once a node has detected that a change may have occurred, there are 2075 three possible cases: 2077 1. Only a change in the Response direction is indicated. There is 2078 nothing that can be done locally; GIMPS must propagate a 2079 notification to its peer. 2081 2. A Query direction change has been detected and a Response 2082 direction change cannot be ruled out. Although some local repair 2083 may be appropriate, it is difficult to decide what, since the 2084 path change may actually have taken place remotely from the 2085 detecting node (so that this node is no longer on the path at 2086 all). 2088 3. A Query direction change has been detected, but there is no 2089 change in the Responding direction. In this case, the detecting 2090 node is the true crossover router, i.e. the point in the network 2091 where old and new paths diverge. It is the correct node to 2092 initiate the local repair process. 2094 In case (3), i.e. at the crossover node, the local repair process is 2095 initiated by the GIMPS level as follows: 2097 o GIMPS marks its routing state information for this flow as 2098 'invalid', unless the route change was actually detected by D-mode 2099 probing (in which case the new state has already been installed). 2101 o GIMPS notifies the local NSLP that local repair is necessary. 2103 It is assumed that the second step will typically trigger the NSLP to 2104 generate a message, and the attempt to send it will stimulate a 2105 GIMPS-Query/Response. This signaling application message will 2106 propagate, also discovering the new route, until it rejoins the old 2107 path; the node where this happens may also have to carry out local 2108 repair actions. 2110 A problem is that there is usually no robust technique to distinguish 2111 case (2) from case (3), because of the relative weakness of the 2112 techniques in determining that such changes have not occurred. (They 2113 can be effective in determining that a change has occurred; however, 2114 even where they can tell that the route from the peer has not 2115 changed, they cannot rule out a change beyond that peer.) There is 2116 therefore a danger that multiple nodes within the network would 2117 attempt to carry out local repair in parallel. 2119 One possible technique to address this problem is that a GIMPS node 2120 that detects case (3) locally, rather than initiating local repair 2121 immediately, still sends a route change notification, just in case 2122 (2) actually applies. If the peer locally detects no downstream 2123 route change, it can signal this in the Query direction (e.g. by 2124 setting another flag in the Common-Header of a GIMPS message). This 2125 acts to damp the possibility of a 'local repair storm', at the cost 2126 of an additional peer-peer round trip time. 2128 6.1.4 Local Signaling Application State Removal 2130 After a route change, a signaling application may wish to remove 2131 state at another node which is no longer on the path. However, since 2132 it is no longer on the path, in principle GIMPS can no longer send 2133 messages to it. (In general, provided this state is soft, it will 2134 time out anyway; however, the timeouts involved may have been set to 2135 be very long to reduce signaling load.) The requirement to remove 2136 state in a specific peer node is identified in [23]. 2138 This requirement can be met provided that GIMPS is able to 'remember' 2139 the old path to the signaling application peer for the period while 2140 the NSLP wishes to be able to use it. Since NSLP peers are a single 2141 GIMPS hop apart, the necessary information is just the old entry in 2142 the node's routing state table for that flow. Rather than requiring 2143 the GIMPS level to maintain multiple generations of this information, 2144 it can just be provided to the signaling application in the same node 2145 (in an opaque form), which can store it if necessary and provide it 2146 back to the GIMPS layer in case it needs to be used. This 2147 information is denoted as 'SII-Handle' in the abstract API of 2148 Appendix D; however, the details are an implementation issue which do 2149 not affect the rest of the protocol. 2151 6.1.5 Operation with Heterogeneous NSLPs 2153 A potential problem with route change detection is that the detecting 2154 GIMPS node may not implement all the signaling applications that need 2155 to be informed. Therefore, it would need to be able to send a 2156 notification back along the unchanged path to trigger the nearest 2157 signaling application aware node to take action. If multiple 2158 signaling applications are in use, it would be hard to define when to 2159 stop propagating this notification. However, given the rules on 2160 message interception and routing state maintenance in Section 4.3, 2161 Section 4.4 and Section 5.3.2.1, this situation cannot arise: all 2162 NSLP peers are exactly one GIMPS hop apart. 2164 The converse problem is that the ability of GIMPS to detect route 2165 changes by purely local monitoring of forwarding tables is more 2166 limited. (This is probably an appropriate limitation of GIMPS 2167 functionality. If we need a protocol for distributing notifications 2168 about local changes in forwarding table state, a flow signaling 2169 protocol is probably not the right starting point.) 2171 6.2 Policy-Based Forwarding and Flow Wildcarding 2173 Signaling messages almost by definition need to contain address and 2174 port information to identify the flow they are signaling for. We can 2175 divide this information into two categories: 2177 Message-Routing-Information: This is the information needed to 2178 determine how a message is routed within the network. It may 2179 include a number of flow N-tuple parameters, and is carried as an 2180 object in each GIMPS message (see Section 5.1). 2182 Additional Packet Classification Information: This is any further 2183 higher layer information needed to select a subset of packets for 2184 special treatment by the signaling application. The need for this 2185 is highly signaling application specific, and so this information 2186 is invisible to GIMPS (if indeed it exists); it will be carried 2187 only in the corresponding NSLP. 2189 The correct pinning of signaling messages to the data path depends on 2190 how well the downstream messages in datagram mode can be made to be 2191 routed correctly. Two strategies are used: 2193 The messages themselves match the flow in destination address and 2194 possibly other fields (see Section 5.3 and Section 5.3.2 for 2195 further discussion). In many cases, this will cause the messages 2196 to be routed correctly even by GIMPS-unaware nodes. 2198 A GIMPS-aware node carrying out policy based forwarding on higher 2199 layer identifiers (in particular, the protocol and port numbers 2200 for IPv4) should take into account the entire 2201 Message-Routing-Information object in selecting the outgoing 2202 interface rather than relying on the IP layer. 2204 The current Message-Routing-Information format allows a limited 2205 degree of 'wildcarding', for example by applying a prefix length to 2206 the source or destination address, or by leaving certain fields 2207 unspecified. A GIMPS-aware node must verify that all flows matching 2208 the Message-Routing-Information would be routed identically in the 2209 downstream direction, or else reject the message with an error. 2211 6.3 NAT Traversal 2213 As already noted, GIMPS messages must carry packet addressing and 2214 higher layer information as payload data in order to define the flow 2215 signalled for. (This applies to all GIMPS messages, regardless of 2216 how they are encapsulated or which direction they are travelling in.) 2217 At an addressing boundary the data flow packets will have their 2218 headers translated; if the signaling payloads are not likewise 2219 translated, the signaling messages will refer to incorrect (and 2220 probably meaningless) flows after passing through the boundary. In 2221 addition, some GIMPS messages (those used in the discovery process) 2222 carry addressing information about the GIMPS nodes themselves, and 2223 this must also be processed appropriately when traversing a NAT. 2225 The simplest solution to this problem is to require that a NAT is 2226 GIMPS-aware, and to allow it to modify datagram mode messages based 2227 on the contents of the Message-Routing-Information payload. (This is 2228 making the implicit assumption that NATs only rewrite the header 2229 fields included in this payload, and not higher layer identifiers.) 2230 Provided this is done consistently with the data flow header 2231 translation, signaling messages will be valid each side of the 2232 boundary, without requiring the NAT to be signaling application 2233 aware. An outline of the set of operations necessary on a downstream 2234 datagram mode message is as follows: 2236 1. Verify that bindings for the data flow are actually in place. 2238 2. Create bindings for subsequent C-mode signaling (based on the 2239 information in the Node-Addressing field). 2241 3. Create a new Message-Routing-Information payload with fields 2242 modified according to the data flow bindings. 2244 4. Create a new Node-Addressing payload with fields to force 2245 upstream D-mode messages through the NAT, and to allow C-mode 2246 exchanges using the C-mode signaling bindings. 2248 5. Add a new NAT-Traversal payload, listing the objects which have 2249 been modified and including the unmodified 2250 Message-Routing-Information. 2252 6. Forward the message with these new payloads. 2254 The original Message-Routing-Information payload is retained in the 2255 message, but encapsulated in the new TLV type. Further information 2256 can be added corresponding to the Node-Addressing payload, either the 2257 original payload itself or, in the case of a GIMPS node that wished 2258 to do topology hiding, opaque tokens (or it could be omitted 2259 altogether). In the case of a sequence of NATs, this part of the 2260 NAT-Traversal object would become a list. Note that a consequence of 2261 this approach is that the routing state tables at the actual 2262 signaling application peers (either side of the NAT) are no longer 2263 directly compatible. In particular, the values of 2264 Message-Routing-Information are different, which is why the 2265 unmodified MRI is propagated in the NAT-Traversal payload to allow 2266 subsequent C-mode messages to be interpreted correctly.. 2268 The case of traversing a GIMPS-unaware NAT is for further study. 2269 There is a dual problem of whether the GIMPS peers either side of the 2270 boundary can work out how to address each other, and whether they can 2271 work out what translation to apply to the Message-Routing-Information 2272 from what is done to the signaling packet headers. The fundamental 2273 problem is that GIMPS messages contain 3 or 4 interdependent 2274 addresses which all have to be consistently translated, and existing 2275 generic NAT traversal techniques such as STUN [19] can process only 2276 two. 2278 6.4 Interaction with IP Tunnelling 2280 The interaction between GIMPS and IP tunnelling is very simple. An 2281 IP packet carrying a GIMPS message is treated exactly the same as any 2282 other packet with the same source and destination addresses: in other 2283 words, it is given the tunnel encapsulation and forwarded with the 2284 other data packets. 2286 Tunnelled packets will not be identifiable as GIMPS messages until 2287 they leave the tunnel, since any router alert option and the standard 2288 GIMPS protocol encapsulation (e.g. port numbers) will be hidden 2289 behind the standard tunnel header. If signaling is needed for the 2290 tunnel itself, this has to be initiated as a separate signaling 2291 session by one of the tunnel endpoints - that is, the tunnel counts 2292 as a new flow. Because the relationship between signaling for the 2293 'microflow' and signaling for the tunnel as a whole will depend on 2294 the signaling application in question, we are assuming that it is a 2295 signaling application responsibility to be aware of the fact that 2296 tunnelling is taking place and to carry out additional signaling if 2297 necessary; in other words, one tunnel endpoint must be signaling 2298 application aware. 2300 In some cases, it is the tunnel exit point (i.e. the node where 2301 tunnelled data and downstream signaling packets leave the tunnel) 2302 that will wish to carry out the tunnel signaling, but this node will 2303 not have knowledge or control of how the tunnel entry point is 2304 carrying out the data flow encapsulation. This information could be 2305 carried as additional data (an additional GIMPS payload) in the 2306 tunnelled signaling packets if the tunnel entry point was at least 2307 GIMPS-aware. This payload would be the GIMPS equivalent of the RSVP 2308 SESSION_ASSOC object of [11]. Whether this functionality should 2309 really be part of GIMPS and if so how the payload should be handled 2310 will be considered in a later version. 2312 6.5 IPv4-IPv6 Transition and Interworking 2314 GIMPS itself is essentially IP version neutral (version dependencies 2315 are isolated in the formats of the Message-Routing-Information and 2316 Node-Addressing TLVs, and GIMPS also depends on the version 2317 independence of the protocols that support messaging associations). 2318 In mixed environments, GIMPS operation will be influenced by the IP 2319 transition mechanisms in use. This section provides a high level 2320 overview of how GIMPS is affected, considering only the currently 2321 predominant mechanisms. 2323 Dual Stack: (This applies both to the basic approach described in 2324 [24] as well as the dual-stack aspects of more complete 2325 architectures such as [28].) In mixed environments, GIMPS should 2326 use the same IP version as the flow it is signaling for; hosts 2327 which are dual stack for applications and routers which are dual 2328 stack for forwarding should have GIMPS implementations which can 2329 support both IP versions. 2331 In theory, for some connection mode encapsulation options, a 2332 single messaging association could carry signaling messages for 2333 flows of both IP versions, but the saving seems of limited value. 2334 The IP version used in datagram mode is closely tied to the IP 2335 version used by the data flow, so it is intrinsically impossible 2336 for a IPv4-only or IPv6-only GIMPS node to support signaling for 2337 flows using the other IP version. 2339 Applications with a choice of IP versions might select a version 2340 for which GIMPS support was available in the network, which could 2341 be established by running parallel discovery procedures. In 2342 theory, a GIMPS message related to a flow of one IP version could 2343 flag support for the other; however, given that IPv4 and IPv6 2344 could easily be separately routed, the correct GIMPS peer for a 2345 given flow might well depend on IP version anyway, making this 2346 flagged information irrelevant. 2348 Packet Translation: (Applicable to SIIT [5] and NAT-PT [12].) Some 2349 transition mechanisms allow IPv4 and IPv6 nodes to communicate by 2350 placing packet translators between them. From the GIMPS 2351 perspective, this should be treated essentially the same way as 2352 any other NAT operation (e.g. between 'public' and 'private' 2353 addresses) as described in Section 6.3. In other words, the 2354 translating node needs to be GIMPS-aware; it will run GIMPS with 2355 IPv4 on some interfaces and with IPv6 on others, and will have to 2356 translate the Message-Routing-Information payload between IPv4 and 2357 IPv6 formats for flows which cross between the two. The 2358 translation rules for the fields in the payload (including e.g. 2359 traffic class and flow label) are as defined in [5]. 2361 Tunnelling: (Applicable to 6to4 [13] and a whole host of other 2362 tunnelling schemes.) Many transition mechanisms handle the problem 2363 of how an end to end IPv6 (or IPv4) flow can be carried over 2364 intermediate IPv4 (or IPv6) regions by tunnelling; the methods 2365 tend to focus on minimising the tunnel administration overhead. 2367 From the GIMPS perspective, the treatment should be as similar as 2368 possible to any other IP tunnelling mechanism, as described in 2369 Section 6.4. In particular, the end to end flow signaling will 2370 pass transparently through the tunnel, and signaling for the 2371 tunnel itself will have to be managed by the tunnel endpoints. 2372 However, additional considerations may arise because of special 2373 features of the tunnel management procedures. For example, [14] 2374 is based on using an anycast address as the destination tunnel 2375 endpoint. It might be unwise to carry out signaling for the 2376 tunnel to such an address, and the GIMPS implementation there 2377 would not be able to use it as a source address for its own 2378 signaling messages (e.g. GIMPS-responses). Further analysis will 2379 be contained in a future version of this specification. 2381 7. Security Considerations 2383 The security requirement for the GIMPS layer is to protect the 2384 signaling plane against identified security threats. For the 2385 signaling problem as a whole, these threats have been outlined in 2386 [21]; the NSIS framework [20] assigns a subset of the responsibility 2387 to the NTLP. The main issues to be handled can be summarised as: 2389 Message Protection: Signaling message content should be protected 2390 against eavesdropping, modification, injection and replay while in 2391 transit. This applies both to GIMPS payloads, and GIMPS should 2392 also provide such protection as a service to signaling 2393 applications between adjacent peers. 2395 State Integrity Protection: It is important that signaling messages 2396 are delivered to the correct nodes, and nowhere else. Here, 2397 'correct' is defined as 'the appropriate nodes for the signaling 2398 given the Message-Routing-Information'. In the case where the MRI 2399 is the Flow Identification for path-coupled signaling, 2400 'appropriate' means 'the same nodes that the infrastructure will 2401 route data flow packets through'. (GIMPS has no role in deciding 2402 whether the data flow itself is being routed correctly; all it can 2403 do is ensure the signaling is routed consistently with it.) GIMPS 2404 uses internal state to decide how to route signaling messages, and 2405 this state needs to be protected against corruption. 2407 Prevention of Denial of Service Attacks: GIMPS nodes and the network 2408 have finite resources (state storage, processing power, 2409 bandwidth). The protocol should try to minimise exhaustion 2410 attacks against these resources and not allow GIMPS nodes to be 2411 used to launch attacks on other network elements. 2413 The main missing issue is handling authorisation for executing 2414 signaling operations (e.g. allocating resources). This is assumed 2415 to be done in each signaling application. 2417 In many cases, GIMPS relies on the security mechanisms available in 2418 messaging associations to handle these issues, rather than 2419 introducing new security measures. Obviously, this requires the 2420 interaction of these mechanisms with the rest of the GIMPS protocol 2421 to be understood and verified, and some aspects of this are discussed 2422 in Section 5.5. 2424 7.1 Message Confidentiality and Integrity 2426 GIMPS can use messaging association functionality, such as TLS or 2427 IPsec, to ensure message confidentiality and integrity. In many 2428 cases, confidentiality of GIMPS information itself is not likely to 2429 be a prime concern, in particular since messages are often sent to 2430 parties which are unknown ahead of time, although the content visible 2431 even at the GIMPS level gives significant opportunities for traffic 2432 analysis. Signaling applications may have their own mechanism for 2433 securing content as necessary; however, they may find it convenient 2434 to rely on protection provided by messaging associations, since it 2435 runs unbroked between signaling application peers. 2437 7.2 Peer Node Authentication 2439 Cryptographic protection (of confidentiality or integrity) requires a 2440 security association with session keys, which can be established 2441 during an authentication and key exchange protocol run based on 2442 shared secrets, public key techniques or a combination of both. 2443 Authentication and key agreement is possible using the protocols 2444 associated with the messaging association being secured (TLS 2445 incorporates this functionality directly; IKE, IKEv2 or KINK can 2446 provide it for IPsec). GIMPS nodes rely on these protocols to 2447 authenticate the identity of the next hop, and GIMPS has no 2448 authentication capability of its own. 2450 However, with discovery, there are few effective ways to know what is 2451 the legitimate next or previous hop as opposed to an impostor. In 2452 other words, cryptographic authentication here only provides 2453 assurance that a node is 'who' it is (i.e. the legitimate owner of 2454 identity in some namespace), not 'what' it is (i.e. a node which is 2455 genuinely on the flow path and therefore can carry out signaling for 2456 a particular flow). Authentication provides only limited protection, 2457 in that a known peer is unlikely to lie about its role. Additional 2458 methods of protection against this type of attack are considered in 2459 Section 7.3 below. 2461 It is open whether peer node authentication should be made signaling 2462 application dependent; for example, whether successful authentication 2463 could be made dependent on presenting authorisation to act in a 2464 particular signaling role (e.g. signaling for QoS). The abstract 2465 API of Appendix D allows GIMPS to forward such policy and 2466 authentication decisions to the NSLP it is serving. 2468 7.3 Routing State Integrity 2470 The internal state in a node (see Section 4.2), specifically the peer 2471 identification, is used to route messages. If this state is 2472 corrupted, signaling messages may be misdirected. 2474 In the case where the message routing method is path-coupled 2475 signaling, the messages need to be routed identically to the data 2476 flow described by the Flow Identifier, and the routing state table is 2477 the GIMPS view of how these flows are being routed through the 2478 network in the immediate neighbourhood of the node. Routes are only 2479 weakly secured (e.g. there is usually no cryptographic binding of a 2480 flow to a route), and there is no other authoritative information 2481 about flow routes than the current state of the network itself. 2482 Therefore, consistency between GIMPS and network routing state has to 2483 be ensured by directly interacting with the routing mechanisms to 2484 ensure that the signaling peers are the appropriate ones for any 2485 given flow. A good overview of security issues and techniques in 2486 this sort of context is provided in [27]. 2488 In one direction, peer identification is installed and refreshed only 2489 on receiving a GIMPS-Reponse message (compare Figure 4). This must 2490 echo the cookie from a previous GIMPS-Query message, which will have 2491 been sent along the flow path (in datagram mode, i.e. end-to-end 2492 addressed). Hence, only the true next peer or an on-path attacker 2493 will be able to generate such a message, provided freshness of the 2494 cookie can be checked at the querying node. 2496 In the reverse direction, peer identification can be installed 2497 directly on receiving a GIMPS-Query message containing addressing 2498 information for the signaling source. However, any node in the 2499 network could generate such a message (indeed, almost any node in the 2500 network could be the genuine upstream peer for a given flow). To 2501 protect against this, two strategies are possible: 2503 Filtering: the receiving node may be able to reject signaling 2504 messages which claim to be for flows with flow source addresses 2505 which would be ruled out by ingress filtering. An extension of 2506 this technique would be for the receiving node to monitor the data 2507 plane and to check explicitly that the flow packets are arriving 2508 over the same interface and if possible from the same link layer 2509 neighbour as the datagram mode signaling packets. (If they are 2510 not, it is likely that at least one of the signaling or flow 2511 packets is being spoofed.) Signaling applications should only 2512 install state on the route taken by the signaling itself. 2514 Authentication (weak or strong): the receiving node may refuse to 2515 install upstream state until it has completed a GIMPS-Confirm 2516 handshaked with the peer. This echoes the response cookie of the 2517 GIMPS-Response, and discourages nodes from using forged source 2518 addresses. A stronger approach is to require full peer 2519 authentication within the messaging association, the reasoning 2520 being that an authenticated peer can be trusted not to pretend 2521 that it is on path when it is not. 2523 The second technique also plays a role in denial of service 2524 prevention, see below. In practice, a combination of both techniques 2525 may be appropriate. 2527 7.4 Denial of Service Prevention 2529 GIMPS is designed so that in general each Query message only 2530 generates at most one Response, so that a GIMPS node cannot become 2531 the source of a denial of service amplification attack. (There is a 2532 special case of retransmitted Response messages, see Section 5.3.3.) 2534 However, GIMPS can still be subjected to denial-of-service attacks 2535 where an attacker using forged source addresses forces a node to 2536 establish state without return routability, causing a problem similar 2537 to TCP SYN flood attacks. In addition to vulnerabilities of a next 2538 peer discovery an unprotected path discovery procedure might 2539 introduce more denial of service attacks since a number of nodes 2540 could possibly be forced to allocate state. Furthermore, an 2541 adversary might modify or replay unprotected signaling messages. 2542 There are two types of state attacks and one computational resource 2543 attack. In the first state attack, an attacker floods a node with 2544 messages that the node has to store until it can determine the next 2545 hop. If the destination address is chosen so that there is no 2546 GIMPS-capable next hop, the node would accumulate messages for 2547 several seconds until the discovery retransmission attempt times out. 2548 The second type of state-based attack causes GIMPS state to be 2549 established by bogus messages. A related 2550 computational/network-resource attack uses unverified messages to 2551 cause a node to make AAA queries or attempt to cryptographically 2552 verify a digital signature. (RSVP is vulnerable to this type of 2553 attack.) Relying only on upper layer security, for example based on 2554 CMS, might open a larger door for denial of service attacks since the 2555 messages are often only one-shot-messages without utilizing multiple 2556 roundtrips and DoS protection mechanisms. 2558 We use a combination of two defences against these attacks: 2560 1. The responding node does not establish a session or discover its 2561 next hop on receiving the GIMPS-Query message, but can wait for a 2562 Confirm message on a secure channel. If the channel exists, the 2563 additional delay is a one one-way delay and the total is no more 2564 than the minimal theoretically possible delay of a three-way 2565 handshake, i.e., 1.5 node-to-node round-trip times. The delay 2566 gets significantly larger if a new connection needs to be 2567 established first. 2569 2. The Response to the Query message contains a cookie. The 2570 previous hop repeats the cookie in the Confirm. State is only 2571 established for messages that contain a valid cookie. The setup 2572 delay is also 1.5 round-trip times. (This mechanism is similar 2573 to that in SCTP [6] and other modern protocols.) 2575 Once a node has decided to establish routing state, there may still 2576 be transport and security state to be established between peers. 2577 This state setup is also vulnerable to additional denial of service 2578 attacks. GIMPS relies on the lower layer protocols that make up 2579 messaging associations to mitigate such attacks. The current 2580 description assumes that the querying node is always the one wishing 2581 to establish a messaging association, so it is typically the 2582 responding node that needs to be protected. 2584 8. IANA Considerations 2586 This section outlines the content of a future IANA considerations 2587 section. 2589 The GIMPS specification requires the creation of registries, as 2590 follows: 2592 GIMPS Message Type: The GIMPS common header (Appendix C.2) contains a 2593 1 byte message type field (initially distinguishing Query, 2594 Response, Confirm and Data messages). 2596 NSLP Identifiers: Each signaling application requires one of more 2597 NSLPIDs (different NSLPIDs may be used to distinguish different 2598 classes of signaling node, for example to handle different 2599 aggregation levels or different processing subsets). An NSLPID 2600 must be associated with a unique RAO value; further considerations 2601 are discussed in Section 5.3.2.1. 2603 Object Types: There is an TBD-bit field in the generic object header 2604 (Appendix C.3.1). Distinguish different ranges for different 2605 allocation styles (standards action, expert review etc.) and 2606 different applicability scopes (experimental/private, 2607 NSLP-specific); by default, object types are public and shared 2608 between all NSLPs. When a new object type is defined, the 2609 extensibility bits (A/B, see Appendix C.3.2) must also be defined. 2611 Extensibility Flags: There are TBD reserved flag bits in the generic 2612 object header (Appendix C.3.1). These are reserved for the 2613 definition of more complex extensibility encoding schemes. 2615 Message Routing Methods: GIMPS allows the idea of multiple message 2616 routing methods (see Section 9.2). The message routing method is 2617 indicated in the leading 2 bytes of the MRI object 2618 (Appendix C.4.1). 2620 Protocol Indicators: The GIMPS design allows the set of possible 2621 protocols to be used in a messaging association to be extended, as 2622 discussed in Section 5.5. Every new mode of using a protocol is 2623 given a single byte Protocol Indicator, which is used as a tag in 2624 the Node Addressing and Stack Proposal objects (Appendix C.4.3 and 2625 Appendix C.4.4). Allocating a new protocol indicator requires 2626 defining the higher layer addressing information (if any) in the 2627 Node Addressing Object that is needed to define its configuration. 2629 Error Classes: There is a 1 byte field at the start of the Value 2630 field of the generic Error object (Appendix C.5.1). Five values 2631 for this field have already been defined. Further general classes 2632 of error could be defined. Note that the value here is primarily 2633 to aid human or management interpretation of otherwise unknown 2634 error codes. 2636 Error Codes: There is a 3 byte error code in the Value field of the 2637 generic Error object (Appendix C.5.1). Error codes are shared 2638 across all NSLPs. When a new error code is allocated, the Error 2639 Class and the format of any associated error-specific information 2640 must also be defined. 2642 9. Open Issues 2644 Note that this section is now partially historic; the authoritative 2645 list of open issues is contained in an online issue tracker at 2646 http://nsis.srmr.co.uk/cgi-bin/roundup.cgi/nsis-ntlp-issues/index. 2647 The subsections remaining here are preserved to keep cross-reference 2648 integrity with the rest of the specification until the issues are 2649 resolved. 2651 9.1 Additional Discovery Mechanisms 2653 The routing state maintenance procedures described in Section 4.4 are 2654 strongly focussed on the problem of discovering, implicitly or 2655 explicitly, the neighbouring peers on the flow path - which is the 2656 necessary functionality for path-coupled signaling. 2658 As well as the GIMPS-Query/Response discovery mechanism for 2659 determining the downstream peer for the path-coupled message routing 2660 method, other techniques may sometimes also be possible. For 2661 example, in many environments, a host has a single access router, 2662 i.e. the downstream peer (for outgoing flows) and the upstream peer 2663 (for incoming ones) are known a priori. More generally, a link state 2664 routing protocol database can be analysed to determine downstream 2665 peers in more complex topologies, and maybe upstream ones if strict 2666 ingress filtering is in effect. More radically, much of the GIMPS 2667 protocol is unchanged if we consider off-path signaling nodes, 2668 although there are significant differences in some of the security 2669 analysis (Section 7.3). None of these possibilities are currently 2670 considered further in this specification. However, the basic 2671 protocol description is unchanged if an encapsulation mechanism is 2672 defined for sending Query messages upstream or directed to particular 2673 nodes, if this information is available from other sources. 2675 9.2 Alternative Message Routing Requirements 2677 The initial assumption of GIMPS is that signaling messages are to be 2678 routed identically to data flow messages. For this case of 2679 path-coupled signaling, the MRI and upstream/downstream flag (in the 2680 Common-Header) define the flow and the relationship of the signaling 2681 to it sufficiently for GIMPS to route its messages correctly. 2682 However, some additional modes of routing signaling messages have 2683 been identified: 2685 Predictive Routing: Here, the intent is to send signaling along a 2686 path that the data flow may or will follow in the future. 2687 Possible cases are pre-installation of state on the backup path 2688 that would be used in the event of a link failure; and predictive 2689 installation of state on the path that will be used after a mobile 2690 node handover. It is currently unclear whether these cases can be 2691 met using the existing GIMPS routing capabilities (and if they 2692 cannot, whether they are in the initial scope of the work). 2694 NAT Address Reservations: This applies to the case where a node 2695 behind a NAT wishes to use NSIS signaling to reserve an address 2696 from which it can be reached by a sender on the other side. This 2697 requires a message to be sent outbound from what will be the flow 2698 receiver although no reverse routing state exists. A possible 2699 solution is described in [29], where the Query is sent towards a 2700 configured address in the 'public' Internet, and intercepted at 2701 the private network boundary. 2703 In the current structure of the protocol definition, the way to 2704 handle these requirements (if they are needed) is to define a new 2705 message routing method which replaces the basic path-coupled version. 2706 The requirements for defining a new routing method include the 2707 following: 2709 o Defining the format of the MRI for the new message routing method 2710 type. 2712 o Defining how Query messages should be encapsulated and routed 2713 corresponding to this MRI. 2715 o Defining any filtering or other security mechanisms that should be 2716 used to validate the MRI in a message. 2718 o Defining how the MRI format is processed on passing through a NAT. 2720 9.3 Message Format Issues 2722 NSIS message formats are defined as a set of objects (see 2723 Appendix C.1). Some aspects are left open: 2725 Ordering: Traditionally, Internet protocols require a node to be able 2726 to process a message with objects in any order. However, this has 2727 some costs in parser complexity, testing interoperability, ease of 2728 compression; there is a special issue with GIMPS that for 2729 efficiency, the NSLP-Data object (which may be large) should come 2730 last. Should object order be fixed or unspecified? 2732 NSLP Versioning: The current working assumption is that if an NSLP 2733 for a particular signaling application is changed so radically 2734 that it is no longer backwards compatible, an entirely new NSLPID 2735 will be allocated. However, this leads to a problem when a node 2736 supporting both variants needs to discover its downstream peer. 2738 If it probes for the 'early' NSLPID it will not detect the case 2739 where the downstream peer supports the later one; if it probes for 2740 the 'later' NSLPID, a downstream peer supporting only the early 2741 variant will bypass the message altogether. The implication is 2742 that a single NSLPID should be used even in this case, with 2743 demultiplexing based on a separate version number (which could be 2744 carried in the common header, or within the NSLP payload). 2746 10. Change History 2748 10.1 Changes In Version -05 2750 Version -05 reformulates the specification, to describe routing state 2751 maintenance in terms of exchanging explicitly identified 2752 Query/Response/Confirm messages, leaving the upstream/downstream 2753 distinction as a specific detail of how Query messages are 2754 encapsulated. This necessitated widespread changes in the 2755 specification text, especially Section 4.2.1, Section 4.4, 2756 Section 5.1 and Section 5.3 (although the actual message sequences 2757 are unchanged). A number of other issues, especially in the area of 2758 message encapsulation, have also been closed. The main changes are 2759 the following: 2761 1. Added a reference to [29] as a concrete example of an alternative 2762 message routing method. 2764 2. Added further text (particularly in Section 2) on what GIMPS 2765 means by the concept of 'session'. 2767 3. Firmed up the selection of UDP as the encapsulation choice for 2768 datagram mode, removing the open issue on this topic. 2770 4. Defined the interaction between GIMPS and signaling applications 2771 for communicating about the cryptographic security properties of 2772 how a message will be sent or has been received (see 2773 Section 4.1.2 and Appendix D). 2775 5. Closed the issue on whether Query messages should use the 2776 signaling or flow source address in the IP header; both options 2777 are allowed by local policy and a flag in the common header 2778 indicates which was used. (See Section 5.3.2.2.) 2780 6. Added the necessary information elements to allow the IP hop 2781 count between adjacent GIMPS peers to be measures and reported. 2782 (See Section 5.2.2 and Appendix C.4.3.) 2784 7. The old open-issue text on selection of IP router alert option 2785 values has been moved into the main specification to capture the 2786 technical considerations that should be used in assigning such 2787 values (in section Section 5.3.2.1). 2789 8. Resolved the open issue on lost Confirm messages by allowing a 2790 choice of timer-based retransmission of the Response, or an 2791 error message from the responding node which causes the 2792 retransmission of the Confirm (see Section 5.3.3). 2794 9. Closed the open issue on support for message scoping (this is now 2795 assumed to be a NSLP function). 2797 10. Moved the authoritative text for most of the remaining open 2798 issues in Section 9 to an online issue tracker. 2800 10.2 Changes In Version -04 2802 Version -04 includes mainly clarifications of detail and extensions 2803 in particular technical areas, in part to support ongoing 2804 implementation work. The main details are as follows: 2806 1. Substantially updated Section 4, in particular clarifying the 2807 rules on what messages are sent when and with what payloads 2808 during routing and messaging association setup, and also adding 2809 some further text on message transfer attributes. 2811 2. The description of messaging association protocol negotiation 2812 including the related object formats has been centralised in a 2813 new Section 5.5, removing the old Section 6.6 and also closing 2814 old open issues 8.5 and 8.6. 2816 3. Made a number of detailed changes in the message format 2817 definitions (Appendix C), as well as incorporating initial rules 2818 for encoding message extensibility information. Also included 2819 explicit formats for a general purpose Error object, and the 2820 objects used to negotiate messaging association protocols. 2821 Updated the corresponding open issues section (Section 9.3) with 2822 a new item on NSLP versioning. 2824 4. Updated the GIMPS API (Appendix D), including more precision on 2825 message transfer attributes, making the NSLP hint about storing 2826 reverse path state a return value rather than a separate 2827 primitive, and adding a new primitive to allow signaling 2828 applications to invalidate GIMPS routing state. Also, added a 2829 new parameter to SendMessage to allow signaling applications to 2830 'bypass' a message statelessly, preserving the source of an 2831 input message. 2833 5. Added an outline for the future content of an IANA considerations 2834 section (Section 8). Currently, this is restricted to 2835 identifying the registries and allocations required, without 2836 defining the allocation policies and other considerations 2837 involved. 2839 6. Shortened the background design discussion in Section 3. 2841 7. Made some clarifications in the terminology section relating to 2842 how the use of C-mode does and does not mandate the use of 2843 transport or security protection. 2845 8. The ABNF for message formats in Section 5.1 has been re-written 2846 with a grammar structured around message purpose rather than 2847 message direction, and additional explanation added to the 2848 information element descriptions in Section 5.2. 2850 9. The description of the datagram mode transport in Section 5.3 has 2851 been updated. The encapsulation rules (covering IP addressing 2852 and UDP port allocation) have been corrected, and a new 2853 subsection on message retransmission and rate limiting has been 2854 added, superceding the old open issue on the same subject 2855 (section 8.10). 2857 10. A new open issue on IP TTL measurement to detect non-GIMPS 2858 capable hops has been added (old section 9.5). 2860 10.3 Changes In Version -03 2862 Version -03 includes a number of minor clarifications and extensions 2863 compared to version -02, including more details of the GIMPS API and 2864 messaging association setup and the node addressing object. The full 2865 list of changes is as follows: 2867 1. Added a new section pinning down more formally the interaction 2868 between GIMPS and signaling applications (Section 4.1), in 2869 particular the message transfer attributes that signaling 2870 applications can use to control GIMPS (Section 4.1.2). 2872 2. Added a new open issue identifying where the interaction between 2873 the security properties of GIMPS and the security requirements of 2874 signaling applications should be identified (old section 9.10). 2876 3. Added some more text in Section 4.2.1 to clarify that GIMPS has 2877 the (sole) responsibility for generating the messages that 2878 refresh message routing state. 2880 4. Added more clarifying text and table to GHC and IP TTL handling 2881 discussion of Section 4.3.4. 2883 5. Split Section 4.4 into subsections for different scenarios, and 2884 added more detail on Node-Addressing object content and use to 2885 handle the case where association re-use is possible in 2886 Section 4.4.2. 2888 6. Added strawman object formats for Node-Addressing and 2889 Stack-Proposal objects in Section 5.1 and Appendix C. 2891 7. Added more detail on the bundling possibilities and appropriate 2892 configurations for various transport protocols in Section 5.4.1. 2894 8. Included some more details on NAT traversal in Section 6.3, 2895 including a new object to carry the untranslated address-bearing 2896 payloads, the NAT-Traversal object. 2898 9. Expanded the open issue discussion in Section 9.3 to include an 2899 outline set of extensibility flags. 2901 10.4 Changes In Version -02 2903 Version -02 does not represent any radical change in design or 2904 structure from version -01; the emphasis has been on adding details 2905 in some specific areas and incorporation of comments, including early 2906 review comments. The full list of changes is as follows: 2908 1. Added a new Section 1.1 which summarises restrictions on scope 2909 and applicability; some corresponding changes in terminology in 2910 Section 2. 2912 2. Closed the open issue on including explicit GIMPS state teardown 2913 functionality. On balance, it seems that the difficulty of 2914 specifying this correctly (especially taking account of the 2915 security issues in all scenarios) is not matched by the saving 2916 of state enabled. 2918 3. Removed the option of a special class of message transfer for 2919 reliable delivery of a single message. This can be implemented 2920 (inefficiently) as a degenerate case of C-mode if required. 2922 4. Extended Appendix C with a general discussion of rules for 2923 message and object formats across GIMPS and other NSLPs. Some 2924 remaining open issues are noted in Section 9.3. 2926 5. Updated the discussion of Section 5.3.2.1 to take into account 2927 the proposed message formats and rules for allocation of NSLP 2928 id, and propose considerations for allocation of RAO values. 2930 6. Modified the description of the information used to route 2931 messages (first given in Section 4.2.1 but also throughout the 2932 document). Previously this was related directly to the flow 2933 identification and described as the Flow-Routing-Information. 2934 Now, this has been renamed Message-Routing-Information, and 2935 identifies a message routing method and any associated 2936 addressing. 2938 7. Modified the text in Section 4.3 and elsewhere to impose sanity 2939 checks on the Message-Routing-Information carried in C-mode 2940 messages, including the case where these messages are part of a 2941 GIMPS-Query/Response exchange. 2943 8. Added rules for message forwarding to prevent message looping in 2944 a new Section 4.3.4, including rules on IP TTL and GIMPS hop 2945 count processing. These take into account the new RAO 2946 considerations of Section 5.3.2.1. 2948 9. Added an outline mechanism for messaging association protocol 2949 stack negotiation, with the details in a new Section 6.6 and 2950 other changes in Section 4.4 and the various sections on message 2951 formats. 2953 10. Removed the open issue on whether storing reverse routing state 2954 is mandatory or optional. This is now explicit in the API 2955 (under the control of the local NSLP). 2957 11. Added an informative annex describing an abstract API between 2958 GIMPS and NSLPs in Appendix D. 2960 10.5 Changes In Version -01 2962 The major change in version -01 is the elimination of 2963 'intermediaries', i.e. imposing the constraint that signaling 2964 application peers are also GIMPS peers. This has the consequence 2965 that if a signaling application wishes to use two classes of 2966 signaling transport for a given flow, maybe reaching different 2967 subsets of nodes, it must do so by running different signaling 2968 sessions; and it also means that signaling adaptations for passing 2969 through NATs which are not signaling application aware must be 2970 carried out in datagram mode. On the other hand, it allows the 2971 elimination of significant complexity in the connection mode handling 2972 and also various other protocol features (such as general route 2973 recording). 2975 The full set of changes is as follows: 2977 1. Added a worked example in Section 3.2. 2979 2. Stated that nodes which do not implement the signaling 2980 application should bypass the message (Section 4.3). 2982 3. Decoupled the state handling logic for routing state and 2983 messaging association state in Section 4.4. Also, allow 2984 messaging associations to be used immediately in both directions 2985 once they are opened. 2987 4. Added simple ABNF for the various GIMPS message types in a new 2988 Section 5.1, and more details of the common header and each 2989 object in Section 5.2, including bit formats in Appendix C. The 2990 common header format means that the encapsulation is now the 2991 same for all transport types (Section 5.4.1). 2993 5. Added some further details on datagram mode encapsulation in 2994 Section 5.3, including more explanation of why a well known port 2995 is needed. 2997 6. Removed the possibility for fragmentation over DCCP 2998 (Section 5.4.1), mainly in the interests of simplicity and loss 2999 amplification. 3001 7. Removed all the tunnel mode encapsulations (old sections 5.3.3 3002 and 5.3.4). 3004 8. Fully re-wrote the route change handling description 3005 (Section 6.1), including some additional detection mechanisms 3006 and more clearly distinguishing between upstream and downstream 3007 route changes. Included further details on GIMPS/NSLP 3008 interactions, including where notifications are delivered and 3009 how local repair storms could be avoided. Removed old 3010 discussion of propagating notifications through signaling 3011 application unaware nodes (since these are now bypassed 3012 automatically). Added discussion on how to route messages for 3013 local state removal on the old path. 3015 9. Revised discussion of policy-based forwarding (Section 6.2) to 3016 account for actual FLow-Routing-Information definition, and also 3017 how wildcarding should be allowed and handled. 3019 10. Removed old route recording section (old Section 6.3). 3021 11. Extended the discussion of NAT handling (Section 6.3) with an 3022 extended outline on processing rules at a GIMPS-aware NAT and a 3023 pointer to implications for C-mode processing and state 3024 management. 3026 12. Clarified the definition of 'correct routing' of signaling 3027 messages in Section 7 and GIMPS role in enforcing this. Also, 3028 opened the possibility that peer node authentication could be 3029 signaling application dependent. 3031 13. Removed old open issues on Connection Mode Encapsulation 3032 (section 8.7); added new open issues on Message Routing 3033 (Section 9.2) and Datagram Mode congestion control. 3035 14. Added this change history. 3037 11. References 3039 11.1 Normative References 3041 [1] Katz, D., "IP Router Alert Option", RFC 2113, February 1997. 3043 [2] Bradner, S., "Key words for use in RFCs to Indicate Requirement 3044 Levels", BCP 14, RFC 2119, March 1997. 3046 [3] Crocker, D., Ed. and P. Overell, "Augmented BNF for Syntax 3047 Specifications: ABNF", RFC 2234, November 1997. 3049 [4] Partridge, C. and A. Jackson, "IPv6 Router Alert Option", 3050 RFC 2711, October 1999. 3052 [5] Nordmark, E., "Stateless IP/ICMP Translation Algorithm (SIIT)", 3053 RFC 2765, February 2000. 3055 [6] Stewart, R., Xie, Q., Morneault, K., Sharp, C., Schwarzbauer, 3056 H., Taylor, T., Rytina, I., Kalla, M., Zhang, L. and V. Paxson, 3057 "Stream Control Transmission Protocol", RFC 2960, October 2000. 3059 [7] Kohler, E., "Datagram Congestion Control Protocol (DCCP)", 3060 Internet-Draft draft-ietf-dccp-spec-09, November 2004. 3062 [8] Conta, A., "Internet Control Message Protocol (ICMPv6)for the 3063 Internet Protocol Version 6 (IPv6) Specification", 3064 Internet-Draft draft-ietf-ipngwg-icmp-v3-06, November 2004. 3066 11.2 Informative References 3068 [9] Braden, B., Zhang, L., Berson, S., Herzog, S. and S. Jamin, 3069 "Resource ReSerVation Protocol (RSVP) -- Version 1 Functional 3070 Specification", RFC 2205, September 1997. 3072 [10] Harkins, D. and D. Carrel, "The Internet Key Exchange (IKE)", 3073 RFC 2409, November 1998. 3075 [11] Terzis, A., Krawczyk, J., Wroclawski, J. and L. Zhang, "RSVP 3076 Operation Over IP Tunnels", RFC 2746, January 2000. 3078 [12] Tsirtsis, G. and P. Srisuresh, "Network Address Translation - 3079 Protocol Translation (NAT-PT)", RFC 2766, February 2000. 3081 [13] Carpenter, B. and K. Moore, "Connection of IPv6 Domains via 3082 IPv4 Clouds", RFC 3056, February 2001. 3084 [14] Huitema, C., "An Anycast Prefix for 6to4 Relay Routers", 3085 RFC 3068, June 2001. 3087 [15] Baker, F., Iturralde, C., Le Faucheur, F. and B. Davie, 3088 "Aggregation of RSVP for IPv4 and IPv6 Reservations", RFC 3175, 3089 September 2001. 3091 [16] Rosenberg, J., Schulzrinne, H., Camarillo, G., Johnston, A., 3092 Peterson, J., Sparks, R., Handley, M. and E. Schooler, "SIP: 3093 Session Initiation Protocol", RFC 3261, June 2002. 3095 [17] Price, R., Bormann, C., Christoffersson, J., Hannu, H., Liu, Z. 3096 and J. Rosenberg, "Signaling Compression (SigComp)", RFC 3320, 3097 January 2003. 3099 [18] Arkko, J., Torvinen, V., Camarillo, G., Niemi, A. and T. 3100 Haukka, "Security Mechanism Agreement for the Session 3101 Initiation Protocol (SIP)", RFC 3329, January 2003. 3103 [19] Rosenberg, J., Weinberger, J., Huitema, C. and R. Mahy, "STUN - 3104 Simple Traversal of User Datagram Protocol (UDP) Through 3105 Network Address Translators (NATs)", RFC 3489, March 2003. 3107 [20] Hancock, R., "Next Steps in Signaling: Framework", 3108 Internet-Draft draft-ietf-nsis-fw-07, December 2004. 3110 [21] Tschofenig, H. and D. Kroeselberg, "Security Threats for NSIS", 3111 Internet-Draft draft-ietf-nsis-threats-06, October 2004. 3113 [22] Stiemerling, M., "A NAT/Firewall NSIS Signaling Layer Protocol 3114 (NSLP)", Internet-Draft draft-ietf-nsis-nslp-natfw-04, October 3115 2004. 3117 [23] Bosch, S., Karagiannis, G. and A. McDonald, "NSLP for 3118 Quality-of-Service signaling", 3119 Internet-Draft draft-ietf-nsis-qos-nslp-05, October 2004. 3121 [24] Nordmark, E. and R. Gilligan, "Basic Transition Mechanisms for 3122 IPv6 Hosts and Routers", 3123 Internet-Draft draft-ietf-v6ops-mech-v2-06, September 2004. 3125 [25] Lonvick, C., "SSH Protocol Architecture", 3126 Internet-Draft draft-ietf-secsh-architecture-21, February 2005. 3128 [26] Moskowitz, R., "Host Identity Protocol", 3129 Internet-Draft draft-ietf-hip-base-01, October 2004. 3131 [27] Nikander, P., "Mobile IP version 6 Route Optimization Security 3132 Design Background", Internet-Draft draft-ietf-mip6-ro-sec-02, 3133 October 2004. 3135 [28] Bound, J., "Dual Stack IPv6 Dominant Transition Mechanism 3136 (DSTM)", Internet-Draft draft-bound-dstm-exp-02, January 2005. 3138 [29] Stiemerling, M., "Loose End Message Routing Method for NATFW 3139 NSLP", Internet-Draft draft-stiemerling-nsis-natfw-mrm-01, 3140 February 2005. 3142 Authors' Addresses 3144 Henning Schulzrinne 3145 Columbia University 3146 Department of Computer Science 3147 450 Computer Science Building 3148 New York, NY 10027 3149 US 3151 Phone: +1 212 939 7042 3152 Email: hgs+nsis@cs.columbia.edu 3153 URI: http://www.cs.columbia.edu 3155 Robert Hancock 3156 Siemens/Roke Manor Research 3157 Old Salisbury Lane 3158 Romsey, Hampshire SO51 0ZN 3159 UK 3161 Email: robert.hancock@roke.co.uk 3162 URI: http://www.roke.co.uk 3164 Appendix A. Acknowledgements 3166 This document is based on the discussions within the IETF NSIS 3167 working group. It has been informed by prior work and formal and 3168 informal inputs from: Cedric Aoun, Attila Bader, Bob Braden, Marcus 3169 Brunner, Pasi Eronen, Xiaoming Fu, Ruediger Geib, Eleanor Hepworth, 3170 Cheng Hong, Georgios Karagiannis, Chris Lang, John Loughney, Allison 3171 Mankin, Jukka Manner, Pete McCann, Andrew McDonald, Glenn Morrow, 3172 Dave Oran, Tom Phelan, Takako Sanda, Charles Shen, Melinda Shore, 3173 Martin Stiemerling, Mike Thomas, Hannes Tschofenig, Sven van den 3174 Bosch, Michael Welzl, and Lars Westberg. In particular, Hannes 3175 Tschofenig provided a detailed set of review comments on the security 3176 section, and Andrew McDonald provided the formal description for the 3177 initial packet formats. Chris Lang's implementation work provided 3178 objective feedback on the clarity and feasibility of the 3179 specification. We look forward to inputs and comments from many more 3180 in the future. 3182 Appendix B. Example Message Routing State Table 3184 Figure 7 shows a signaling scenario for a single flow being managed 3185 by two signaling applications using the path-coupled message routing 3186 method. The flow sender and receiver and one router support both, 3187 two other routers support one each. 3189 A B C D E 3190 +------+ +-----+ +-----+ +-----+ +--------+ 3191 | Flow | +-+ +-+ |NSLP1| |NSLP1| | | | Flow | 3192 |Sender|====|R|====|R|====|NSLP2|====| |====|NSLP2|====|Receiver| 3193 | | +-+ +-+ |GIMPS| |GIMPS| |GIMPS| | | 3194 +------+ +-----+ +-----+ +-----+ +--------+ 3196 ------------------------------>> 3197 Flow Direction 3199 Figure 7: A Signaling Scenario 3201 The routing state table at node B is as follows: 3203 +--------------------+----------+----------+-------------+----------+ 3204 | Message Routing | Session | NSLP ID | Response | Query | 3205 | Information | ID | | Direction | Directio | 3206 | | | | | n | 3207 +--------------------+----------+----------+-------------+----------+ 3208 | Method = Path | 0xABCD | NSLP1 | IP-#A | (null) | 3209 | Coupled; Flow ID = | | | | | 3210 | {IP-#A, IP-#E, | | | | | 3211 | protocol, ports} | | | | | 3212 | | | | | | 3213 | Method = Path | 0x1234 | NSLP2 | IP-#A | Pointer | 3214 | Coupled; Flow ID = | | | | to B-D | 3215 | {IP-#A, IP-#E, | | | | messagin | 3216 | protocol, ports} | | | | g | 3217 | | | | | associa | 3218 | | | | | ti on | 3219 +--------------------+----------+----------+-------------+----------+ 3221 The Response direction state is just the same address for each 3222 application. For the Query direction, NSLP1 only requires datagram 3223 mode messages and so no explicit routing state towards C is needed. 3224 NSLP2 requires a messaging association for its messages towards node 3225 D, and node C does not process NSLP2 at all, so the peer state for 3226 NSLP2 is a pointer to a messaging association that runs directly from 3227 B to D. Note that E is not visible in the state table (except 3228 implicitly in the address in the message routing information); 3229 routing state is stored only for adjacent peers. (In addition to the 3230 peer identification, IP hop counts are stored for each peer where the 3231 state itself if not null; this is not shown in the table.) 3233 Appendix C. Bit-Level Formats 3235 This appendix provides initial formats for the various component 3236 parts of the GIMPS messages defined abstractly in Section 5.2. It 3237 should be noted that these formats are extremely preliminary and 3238 should be expected to change completely several times during the 3239 further development of this specification. 3241 In addition, this appendix includes some general rules for the format 3242 of messages and message objects across all protocols in the NSIS 3243 protocol suite (i.e. the current and future NSLPs as well as GIMPS 3244 itself). The intention of these common rules is to encourage 3245 commonality in implementations, ease of testing and debugging, and 3246 sharing of object definitions across different applications. 3248 C.1 General NSIS Formatting Guidelines 3250 Each NSIS message consists of a header and a sequence of objects. An 3251 NSLP message is one object within a GIMPS message. The GIMPS header 3252 has a specific format, described in more detail in Appendix C.2 3253 below; the NSLP header format is common to all signaling applications 3254 and includes simply a message type (which may be structured into a 3255 type field and some processing flags, depending on the application). 3256 Note that GIMPS provides the message length information and signaling 3257 application identification. 3259 Note that there is no version information at the NSLP level. It is 3260 assumed that minor protocol extensions can be implemented by adding 3261 extra objects (see Appendix C.3.2); if an NSLP has to be extended so 3262 much that backwards compatibity is no longer possible, a new 3263 signaling application identifier is allocated instead. An open issue 3264 on this subject is discussed in Section 9.3. 3266 Every object has the following general format: 3268 o The overall format is Type-Length-Value (in that order). 3270 o By default, assignments for the Type field are common across all 3271 NSIS protocols (i.e. there is a single registry). This is to 3272 facilitate the sharing of common objects across different 3273 signaling applications. The allocation of control flags to define 3274 how unknown types should be handled is also common across 3275 signaling applications; this is discussed in Appendix C.3.2. 3277 o Part of the object type space can be set aside for TLVs which for 3278 some reason should only be used within a single signaling 3279 application, see Section 8. 3281 o Length has the units of 32 bit words, and measures the length of 3282 Value. If there is no Value, Length=0. 3284 o Value is (therefore) a whole number of 32 bit words. If there is 3285 any padding required, the length and location must be defined by 3286 the object-specific format information; objects which contain 3287 variable length (e.g. string) types may need to include 3288 additional length subfields to do so. 3290 o Any part of the object used for padding or defined as reserved 3291 must be set to 0 on transmission and must be ignored on reception. 3293 Error messages are identified by containing an error object (i.e. an 3294 object with Type='Error'). The error object format is given in 3295 Appendix C.5.1; its Value field includes an error class, an error 3296 code, and optionally additional error-specific information. Again, 3297 the error code space is common across all protocols. 3299 C.2 The GIMPS Common Header 3301 This header precedes all GIMPS messages. It has a fixed format, as 3302 shown below. Note that (unlike NSLP messages) the GIMPS header does 3303 include a version number, since allocating new lower layer 3304 identifiers to demultiplex a new GIMPS version will be significantly 3305 harder than allocating a new NSLP identifier. 3307 0 1 2 3 3308 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 3309 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3310 | Version | GIMPS hops | Message length | 3311 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3312 | Signaling Application ID | Type |S| Reserved | 3313 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3315 Message length = the total number of words in the message after 3316 the common header itself 3317 Type = the GIMPS message type (Query, Response, etc.) 3318 S flag = set if the IP source address is the signaling 3319 source address, clear if it was derived from the 3320 MRI 3322 C.3 General Object Characteristics 3324 C.3.1 TLV Header 3326 Each object begins with a fixed header giving the object type and 3327 object length. The bits marked 'A' and 'B' are extensibility flags 3328 which are defined below; the remaining bits marked 'r' are reserved. 3330 0 1 2 3 3331 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 3332 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3333 |A|B|r|r| Type |r|r|r|r| Length | 3334 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3336 C.3.2 Object Extensibility 3338 The leading two bits of the common TLV header are used to signal the 3339 desired treatment for objects whose treatment has not been defined in 3340 the protocol specification in question (i.e. whose Type field is 3341 unknown at the receiver). The following four categories of object 3342 have been identified, and are loosely described here. 3344 AB=00 ("Mandatory"): If the object is not understood, the entire 3345 message containing it must be rejected with an error indication. 3347 AB=01 ("Ignore"): If the object is not understood, it should be 3348 deleted and then the rest of the message processed as usual. 3350 AB=10 ("Forward"): If the object is not understood, it should be 3351 retained unchanged in any message forwarded as a result of message 3352 processing, but not stored locally. 3354 AB=11 ("Refresh"): If the object is not understood, it should be 3355 incorporated into the locally stored signaling application state 3356 for this flow/session, forwarded in any resulting message, and 3357 also used in any refresh or repair message which is generated 3358 locally. 3360 For objects used within the NSLP-Data payload, the precise usage of 3361 these flags must be defined for each signaling application. In 3362 particular, signaling applications must define how to indicate 3363 errors, and what it means to forward or refresh 'state'; they may 3364 also restrict whether particular flag combinations can be used. 3366 C.4 GIMPS Specific TLV Objects 3368 The objects defined in this section are expected to be used mainly by 3369 GIMPS rather than signaling applications. 3371 In the following object diagrams, '//' is used to indicate a variable 3372 sized field and ':' is used to indicate a field that is optionally 3373 present. 3375 C.4.1 Message-Routing-Information 3377 Type: Message-Routing-Information 3379 Length: Variable (depends on message routing method) 3381 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3382 | Message-Routing-Method | | 3383 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + 3384 // Method-specific addressing information (variable) // 3385 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3387 In the case of basic path-coupled routing, the addressing information 3388 takes the following format: 3390 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3391 |IP-Ver |P|T|F|I|A|B|D|Reserved | 3392 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3393 // Source Address // 3394 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3395 // Destination Address // 3396 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3397 | Source Prefix | Dest Prefix | Protocol | Traffic Class | 3398 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3399 : Reserved | Flow Label : 3400 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3401 : SPI : 3402 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3403 : Source Port : Destination Port : 3404 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3406 The flags are: 3407 P - Protocol 3408 T - Traffic Class 3409 F - Flow Label 3410 I - SPI 3411 A - Source Port 3412 B - Destination Port 3413 I/A/B can only be set if P is set. 3414 D - Direction of message relative to MRI 3416 F may only be set if IP-Ver is 6. If F is not set, the entire 32 bit 3417 word for the FLow Label is absent. 3419 If only one of S, D is set, both Port fields are included in the 3420 message. However, the contents of the field are only interpreted if 3421 the corresponding flag is set. If the flag is not set, Port values 3422 will be ignored as part of the flow definition; the MRI matches all 3423 packets regardless of port. If the flag is set and Port=0x0000, the 3424 MRI will apply to a specific port, whose value is not yet known. 3426 C.4.2 Session Identification 3428 Type: Session-Identification 3430 Length: Fixed (TBD 4 32-bit words) 3432 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3433 | | 3434 + + 3435 | | 3436 + Session ID + 3437 | | 3438 + + 3439 | | 3440 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3442 C.4.3 Node Addressing 3444 Type: Node-Addressing 3446 Length: Variable (depends on length of Peer-Identity and number of 3447 higher-layer-protocol fields present) 3449 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3450 | PI-Length | HL-Count | IP-TTL |IP-Ver | Rsvd | 3451 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3452 // Peer Identity // 3453 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3454 // Interface Address // 3455 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3456 // Higher-Layer-Information 1 // 3457 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3458 : : 3459 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3460 // Higher-Layer-Information N // 3461 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3463 PI-Length = the byte length of the Peer-Identity field 3464 (note that the Peer-Identity field itself is padded 3465 to a whole number of words) 3466 HL-Count = the number of higher-layer-information fields 3467 (these contain their own length information); 3468 0 if this object is carried in connection mode 3469 IP-TTL = initial or reported IP-TTL 3470 IP-Ver = the IP version for the Interface-Address field 3472 The higher layer information fields are formatted as follows: 3474 o There is a 1-byte Protocol Indicator, as described in Section 5.5. 3476 o There is a 1-byte length field defining the amount of 3477 configuration data that follows after the length field. 3479 o There is a variable length of configuration data. 3481 o There are 0, 1, 2, or 3 bytes of zero padding to the next word 3482 boundary. 3484 Note that the contents of the configuration data may differ depending 3485 on whether the NAO is in a GIMPS-query or GIMPS-response. 3487 C.4.4 Stack Proposal 3489 Type: Stack-Proposal 3491 Length: Variable (depends on number of profiles and size of each 3492 profile) 3494 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3495 | Prof-Count | Reserved | 3496 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3497 // Profile 1 // 3498 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3499 : : 3500 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3501 // Profile 2 // 3502 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3504 Prof-Count = The number of profiles in the proposal 3506 Each profile is itself a sequence of protocol layers, and the profile 3507 is formatted as a list as follows: 3509 o The first byte is a count of the number of layers in the profile. 3511 o This is followed by a sequence of 1-byte Protocol Indicators as 3512 described in Section 5.5. 3514 o The profile is padded to a word boundary with 0, 1, 2 or 3 zero 3515 bytes. 3517 C.4.5 Query Cookie 3519 Type: Query-Cookie 3521 Length: Variable (selected by querying node) 3523 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3524 | | 3525 // Query Cookie // 3526 | | 3527 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3529 Note that the querying node uses the value of the query cookie in the 3530 GIMPS-response message on an existing messaging association to match 3531 with the corresponding GIMPS-query. This imposes certain uniqueness 3532 requirements on the cookie contents. 3534 C.4.6 Responder Cookie 3536 Type: Responder-Cookie 3538 Length: Variable (selected by responding node) 3540 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3541 | | 3542 // Responder Cookie // 3543 | | 3544 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3546 Note that the responding node uses the value of the responder cookie 3547 in the GIMPS-confirm message to match a new messaging association 3548 with the corresponding GIMPS-query/response exchange. This imposes 3549 certain uniqueness requirements on the cookie contents. 3551 C.4.7 Lifetime 3553 Type: Lifetime 3555 Length: Fixed - 1 32-bit word 3556 Value: Routing state lifetime in seconds 3558 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3559 | Lifetime | 3560 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3562 C.4.8 NAT Traversal 3564 Type: NAT-Traversal 3566 Length: Variable (depends on length of contained fields) 3568 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3569 | MRI-Length | Type-Count | NAT-Count | Reserved | 3570 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3571 // Original Message-Routing-Information // 3572 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3573 // List of translated objects // 3574 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3575 | Length of opaque NAO info. | | 3576 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ // 3577 // NAO information replaced by NAT #1 | 3578 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3579 : : 3580 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3581 | Length of opaque NAO info. | | 3582 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ // 3583 // NAO information replaced by NAT #N | 3584 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3586 MRI-Length = the word length of the included MRI payload 3587 Type-Count = the number of GIMPS payloads translated by the 3588 NAT; the Type numbers are included as a list 3589 (padded with 2 null bytes if necessary) 3590 NAT-Count = the number of NATs traversed by the message, and the 3591 number of opaque NAO-related payloads at the end 3592 of the object 3594 C.4.9 NSLP Data 3596 Type: NSLP-Data 3597 Length: Variable (depends on NSLP) 3599 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3600 | | 3601 // NSLP Data // 3602 | | 3603 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3605 C.5 Generic NSIS TLV Objects 3607 The objects defined in this section are general purpose objects, 3608 which will be used by both GIMPS and signaling applications in 3609 general. 3611 C.5.1 Error Object 3613 Type: Error 3615 Length: Variable (depends on error) 3617 Value: Contains a 1 byte error class and 3 byte error code, an error 3618 source identifier and optionally variable length error-specific 3619 information. 3621 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3622 | Error Class | Error Code | 3623 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3624 | ESI-Length | Reserved | 3625 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3626 // Error Source Identifier // 3627 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3628 // Optional error-specific information // 3629 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3631 The first byte "Error Class" indicates the severity level. The 3632 currently defined severity levels are: 3634 Informational: response data which should not be thought of as 3635 changing the condition of the protocol state machine. 3637 Success: response data which indicates that the message being 3638 responded to has been processed successfully in some sense. 3640 Protocol-Error: the message has been rejected because of a protocol 3641 error (e.g. an error in message format). 3643 Transient-Failure: the message has been rejected because of a 3644 particular local node status which may be transient (i.e. it may 3645 be worthwhile to retry after some delay). 3647 Permanent-Failure: the message has been rejected because of local 3648 node status which will not change without additional out of band 3649 (e.g. management) operations. 3651 Additional error class values are reserved. 3653 The allocation of error classes to particular errors is not precise; 3654 the above descriptions are deliberately informal. Actually error 3655 processing should take into account the specific error in question; 3656 the error class may be useful supporting information (e.g. in 3657 network debugging). 3659 The Error Source Identifier can be generated in an 3660 implementation-specific manner. It is suggested that the same method 3661 is used as for the Peer Identity in the Node Addressing object. 3663 ESI-Length is given in bytes (excluding padding). The Error Source 3664 Identifier MUST be padded to make it a whole number of 32-bit words 3665 in length. The optional additional error-specific information fills 3666 the rest of the object up to the length given in the object header. 3668 The Error object may be carried either at the GIMPS level to indicate 3669 GIMPS errors, or at the NSLP level (inside the NSLP-Data object) to 3670 indicate NSLP errors. However, the format and error code assignments 3671 are common to both uses. 3673 Appendix D. API between GIMPS and NSLP 3675 This appendix provides an initial abstract API between GIMPS and 3676 NSLPs. 3678 This does not constrain implementors, but rather helps clarify the 3679 interface between the different layers of the NSIS protocol suite. 3680 In addition, although some of the data types carry the information 3681 from GIMPS Information Elements, this does not imply that the format 3682 of that data as sent over the API has to be the same. 3684 Conceptually the API has similarities to the UDP sockets API, 3685 particularly that for unconnected UDP sockets. An extension for an 3686 API like that for UDP connected sockets could be considered. In this 3687 case, for example, the only information needed in a SendMessage 3688 primitive would be NSLP-Data, NSLP-Data-Size, and NSLP-Message-Handle 3689 (which can be null). Other information which was persistent for a 3690 group of messages could be configured once for the socket. Such 3691 extensions may make a concrete implementation more scalable and 3692 efficient but do not change the API semantics, and so are not 3693 considered further here. 3695 D.1 SendMessage 3697 This primitive is passed from an NSLP to GIMPS. It is used whenever 3698 the NSLP wants to send a message. 3700 SendMessage ( NSLP-Data, NSLP-Data-Size, NSLP-Message-Handle, 3701 NSLP-Id, Session-ID, MRI, 3702 Source-SII-Handle, Peer-SII-Handle, 3703 Transfer-Attributes, Timeout, IP-TTL ) 3705 The following arguments are mandatory. 3707 NSLP-Data: The NSLP message itself. 3709 NSLP-Data-Size: The length of NSLP-Data. 3711 NSLP-Message-Handle: A handle for this message, that can be used 3712 later by GIMPS to reference it in status reports (in particular, 3713 notification about what security attributes will be used for the 3714 message, or error notifications). A NULL handle may be supplied 3715 if the NSLP is not interested in receiving MessageStatus 3716 notifications for this message. 3718 NSLP-Id: An identifier indicating which NSLP this is. 3720 Session-ID: The NSIS session identifier. Note that it is assumed 3721 that the signaling application provides this to GIMPS rather than 3722 GIMPS providing a value itself; often, this will be a value 3723 associated with an existing session, for example received in an 3724 incoming message. In the case of an entirely new session, a GIMPS 3725 implementation might provide library functionality to generate a 3726 new, cryptographically random SID which is guaranteed not to 3727 collide with any existing session. 3729 MRI: Message routing information for use by GIMPS in determining the 3730 correct next GIMPS hop for this message. It contains, for 3731 example, the flow source/destination addresses and the type of 3732 routing to use for the signaling message. The message routing 3733 information implies the message routing method to be used and also 3734 includes the direction of the message. 3736 The following arguments are optional. 3738 Source-SII-Handle: A handle, previously supplied by GIMPS in 3739 RecvMessage, which indicates that the NSLP wishes to originate the 3740 message as though it came from the identified source (e.g. so 3741 responses will be returned to that source). Will cause an error 3742 if set with a large payload or non-trivial Transfer-Attributes. 3744 Peer-SII-Handle: A handle, previously supplied by GIMPS, to a data 3745 structure (identifying peer addresses and interfaces) that should 3746 be used to explicitly route the message to a particular GIMPS next 3747 hop. If supplied, GIMPS should validate that it is consistent 3748 with the MRI. 3750 Transfer-Attributes: Attributes defining how the message should be 3751 handled (see Section 4.1.2). The following attributes can be 3752 considered: 3754 Reliability: Values 'unreliable' (default) or 'reliable'. 3756 Security: This attribute allows the NSLP to specify what level of 3757 security protection is requested for the message (selected from 3758 'integrity' and 'confidentiality'), and can also be used to 3759 specify what authenticated signaling source and destination 3760 identities should be used to send the message. The 3761 possibilities can be learned by the NSLP from prior 3762 MessageStatus or RecvMessage notifications. If an 3763 NSLP-Message-Handle is provided, GIMPS will inform the NSLP of 3764 what values it has actually chosen for this attribute via a 3765 MessageStatus callback. This might take place either 3766 synchronously (where GIMPS is just selecting from available 3767 messaging associations), or asynchronously (when a new 3768 messaging association needs to be created). 3770 Local Processing: This attribute contains hints from the NSLP 3771 about what local policy should be applied to the message; in 3772 particular, its transmission priority relative to other 3773 messages, or whether GIMPS should attempt to set up or maintain 3774 forward routing state. 3776 Timeout: Length of time GIMPS should attempt to send this message 3777 before indicating an error. 3779 IP-TTL: The value of the IP TTL that should be used when sending this 3780 message. 3782 D.2 RecvMessage 3784 This primitive is passed from GIMPS to an NSLP. It is used whenever 3785 GIMPS receives a message from the network. This primitive can return 3786 a value from the NSLP which indicates whether the NSLP wishes GIMPS 3787 to retain message routing state. 3789 RecvMessage ( NSLP-Data, NSLP-Data-Size, NSLP-Id, Session-ID, MRI, 3790 SII-Handle, Transfer-Attributes, IP-TTL, IP-Distance ) 3792 NSLP-Data: The NSLP message itself (may be empty). 3794 NSLP-Data-Size: The length of NSLP-Data (may be zero). 3796 NSLP-Id: An identifier indicating which NSLP this is message is for. 3798 Session-ID: The NSIS session identifier. 3800 MRI: Message routing information that was used by GIMPS in forwarding 3801 this message. It contains, for example, the flow 3802 source/destination addresses, the type of routing used for the 3803 signaling message, and the direction of the message relative to 3804 the MRI. Implicitly defines the message routing method that was 3805 used. 3807 SII-Handle: A handle to a data structure, identifying peer addresses 3808 and interfaces. Can be used to identify route changes and for 3809 explicit routing to a particular GIMPS next hop. 3811 Transfer-Attributes: The reliability and security attributes that 3812 were associated with the reception of this particular message. 3814 IP-TTL: The value of the IP TTL (or Hop Limit) this message was 3815 received with (if available). 3817 IP-Distance: The number of IP hops from the peer signaling node which 3818 sent this message along the path, or 0 if this information is not 3819 available. 3821 D.3 MessageStatus 3823 This primitive is passed from GIMPS to an NSLP. It is used to notify 3824 the NSLP that a message that it requested to be sent has failed to be 3825 dispatched, or to inform the NSLP about the transfer attributes that 3826 have been selected for the message (specifically, security 3827 attributes). The NSLP can respond to this message with a return code 3828 to abort the sending of the message if the attributes are not 3829 acceptable. 3831 MessageStatus ( NSLP-Message-Handle, Transfer-Attributes, Error-Type ) 3833 NSLP-Message-Handle: A handle for the message provided by the NSLP at 3834 the time of sending. 3836 Transfer-Attributes: The reliability and security attributes that 3837 will be used to transmit this particular message. 3839 Error-Type: Indicates the type of error that occurred. For example, 3840 'no next node found'. 3842 D.4 NetworkNotification 3844 This primitive is passed from GIMPS to an NSLP. It indicates that a 3845 network event of possible interest to the NSLP occurred. 3847 NetworkNotification ( MRI, Network-Notification-Type ) 3849 MRI: Provides the message routing information to which the network 3850 notification applies. 3852 Network-Notification-Type: Indicates the type of event that caused 3853 the notification, e.g. downstream route change, upstream route 3854 change, detection that this is the last node. 3856 D.5 SetStateLifetime 3858 This primitive is passed from an NSLP to GIMPS. It indicates the 3859 lifetime for which the NSLP would like GIMPS to retain its state. It 3860 can also give a hint that the NSLP is no longer interested in the 3861 state. 3863 SetStateLifetime ( MRI, Direction, State-Lifetime ) 3865 MRI: Provides the message routing information to which the network 3866 notification applies. 3868 Direction: A flag indicating whether this relates to state for the 3869 upstream or downstream direction (in relation to the MRI). 3871 State-Lifetime: Indicates the lifetime for which the NSLP wishes 3872 GIMPS to retain its state (may be zero, indicating that the NSLP 3873 has no further interest in the GIMPS state). 3875 D.6 InvalidateRoutingState 3877 This primitive is passed from an NSLP to GIMPS. It indicates that 3878 the NSLP has knowledge that the next signaling hop known to GIMPS may 3879 no longer be valid, either because of changes in the network routing 3880 or the processing capabilities of NSLP nodes. It is an indication to 3881 GIMPS to restart the discovery process. 3883 InvalidateRoutingState ( NSLP-Id, MRI, Direction, Urgency ) 3885 NSLP-Id: The NSLP originating the message. May be null (in which 3886 case the invalidation applies to all signaling applications). 3888 MRI: The flow for which routing state should be invalidated. 3890 Direction: A flag indicating whether this relates to state for the 3891 upstream or downstream direction (in relation to the MRI). 3893 Urgency: A hint to GIMPS as to whether rediscovery should take place 3894 immediately, or only when the next signaling message is ready to 3895 be sent. 3897 Intellectual Property Statement 3899 The IETF takes no position regarding the validity or scope of any 3900 Intellectual Property Rights or other rights that might be claimed to 3901 pertain to the implementation or use of the technology described in 3902 this document or the extent to which any license under such rights 3903 might or might not be available; nor does it represent that it has 3904 made any independent effort to identify any such rights. Information 3905 on the procedures with respect to rights in RFC documents can be 3906 found in BCP 78 and BCP 79. 3908 Copies of IPR disclosures made to the IETF Secretariat and any 3909 assurances of licenses to be made available, or the result of an 3910 attempt made to obtain a general license or permission for the use of 3911 such proprietary rights by implementers or users of this 3912 specification can be obtained from the IETF on-line IPR repository at 3913 http://www.ietf.org/ipr. 3915 The IETF invites any interested party to bring to its attention any 3916 copyrights, patents or patent applications, or other proprietary 3917 rights that may cover technology that may be required to implement 3918 this standard. Please address the information to the IETF at 3919 ietf-ipr@ietf.org. 3921 Disclaimer of Validity 3923 This document and the information contained herein are provided on an 3924 "AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS 3925 OR IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY AND THE INTERNET 3926 ENGINEERING TASK FORCE DISCLAIM ALL WARRANTIES, EXPRESS OR IMPLIED, 3927 INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE 3928 INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED 3929 WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. 3931 Copyright Statement 3933 Copyright (C) The Internet Society (2005). This document is subject 3934 to the rights, licenses and restrictions contained in BCP 78, and 3935 except as set forth therein, the authors retain all their rights. 3937 Acknowledgment 3939 Funding for the RFC Editor function is currently provided by the 3940 Internet Society.