idnits 2.17.1 draft-cheshire-pcp-anycast-02.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- == There are 2 instances of lines with non-RFC6890-compliant IPv4 addresses in the document. If these are example addresses, they should be changed. == There are 1 instance of lines with non-RFC3849-compliant IPv6 addresses in the document. If these are example addresses, they should be changed. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (July 14, 2013) is 3939 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) -- No information found for draft-ietf-ipngwg-dns-discovery - is the name correct? Summary: 0 errors (**), 0 flaws (~~), 3 warnings (==), 2 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 PCP working group S. Cheshire 3 Internet-Draft Apple 4 Intended status: Standards Track July 14, 2013 5 Expires: January 15, 2014 7 PCP Anycast Address 8 draft-cheshire-pcp-anycast-02 10 Abstract 12 The Port Control Protocol Anycast Address enables PCP clients to 13 transmit messages to their closest on-path NAT, Firewall, or other 14 middlebox, without having to learn the IP address of that middlebox 15 via some external channel. 17 Status of this Memo 19 This Internet-Draft is submitted in full conformance with the 20 provisions of BCP 78 and BCP 79. 22 Internet-Drafts are working documents of the Internet Engineering 23 Task Force (IETF). Note that other groups may also distribute 24 working documents as Internet-Drafts. The list of current Internet- 25 Drafts is at http://datatracker.ietf.org/drafts/current/. 27 Internet-Drafts are draft documents valid for a maximum of six months 28 and may be updated, replaced, or obsoleted by other documents at any 29 time. It is inappropriate to use Internet-Drafts as reference 30 material or to cite them other than as "work in progress." 32 This Internet-Draft will expire on January 15, 2014. 34 Copyright Notice 36 Copyright (c) 2013 IETF Trust and the persons identified as the 37 document authors. All rights reserved. 39 This document is subject to BCP 78 and the IETF Trust's Legal 40 Provisions Relating to IETF Documents 41 (http://trustee.ietf.org/license-info) in effect on the date of 42 publication of this document. Please review these documents 43 carefully, as they describe your rights and restrictions with respect 44 to this document. Code Components extracted from this document must 45 include Simplified BSD License text as described in Section 4.e of 46 the Trust Legal Provisions and are provided without warranty as 47 described in the Simplified BSD License. 49 1. Introduction 51 The Port Control Protocol document [RFC6887] specifies the message 52 formats used, but the address to which a client sends its request is 53 either assumed to be the default router (which is appropriate in a 54 typical single-link residential network) or has to be configured 55 otherwise via some external mechanism, such as DHCP. 57 One drawback of relying on external configuration is that it creates 58 an external dependency on another piece of network infrastructure 59 which must be configured with the right address for PCP to work. In 60 some environments the staff managing the DHCP servers may not be the 61 same staff managing the NAT gateways, and in any case, constantly 62 keeping the DHCP server address information up to date as NAT 63 gateways are added, removed, or reconfigured, is burdensome. 65 Another drawback of relying on DHCP for configuration is that one of 66 the target deployment environments for PCP -- 3GPP for mobile 67 telephones -- does not use DHCP. 69 One design option that was considered for Apple's NAT gateways was to 70 have the NAT gateway simply handle and respond to all packets 71 addressed to UDP port 5351, regardless of the destination address in 72 the packet. Since the device is a NAT gateway, it already examines 73 every packet in order to rewrite port numbers, so also detecting 74 packets addressed to UDP port 5351 is not a significant additional 75 burden. Also, since this device is a NAT gateway which rewrites port 76 numbers, any attempt by a client to talk *though* this first NAT 77 gateway to create mappings in some second upstream NAT gateway is 78 futile and pointless. Any mappings created in the second NAT gateway 79 are useful to the client only if there are also corresponding 80 mappings created in the first NAT gateway. Consequently, there is no 81 case where it is useful for PCP requests to pass transparently 82 through the first PCP-aware NAT gateway on their way to the second 83 PCP-aware NAT gateway. In all cases, for useful connectivity to be 84 established, the PCP request must be handled by the first NAT 85 gateway, and then the first NAT gateway generates a corresponding new 86 upstream request to establish a mapping in the second NAT gateway. 87 (This process can be repeated recursively for as many times as 88 necessary for the depth of nesting of NAT gateways; this is 89 transparent to the client device [Recurs].) 91 These two issues result in the following related observations: the 92 PCP client may not *know* what destination address to use in its PCP 93 request packets; the PCP server doesn't *care* what destination 94 address is in the PCP request packets. 96 Given that the devices neither need to know nor care what destination 97 address goes in the packet, all we need to do is pick one and use it. 98 It's little more than a placeholder in the IP header. Any globally 99 routable unicast address will do. Since this address is one that 100 automatically routes its packet to the closest on-path device that 101 implements the desired functionality, it is an anycast address. 103 In the simple case where the first-hop router is also the NAT gateway 104 (as is common in a typical single-link residential network), sending 105 to the PCP anycast address is equivalent to sending to the client's 106 default router, as specified in the PCP base document [RFC6887]. 108 In the case of a larger corporate network, where there may be several 109 internal routed subnets and one or more border NAT gateway(s) 110 connecting to the rest of the Internet, sending to the PCP anycast 111 address has the interesting property that it magically finds the 112 right border NAT gateway for that client. Since we posit that other 113 network infrastructure does not need (and should not have) any 114 special knowledge of PCP (or its anycast address) this means that to 115 other non-NAT routers, the PCP anycast address will look like any 116 other unicast destination address on the public Internet, and 117 consequently the packet will be forwarded as for any other packet 118 destined to the public Internet, until it reaches a NAT or firewall 119 device that is aware of the PCP anycast address. This will result in 120 the packet naturally arriving the NAT gateway that handles this 121 client's outbound traffic destined to the public Internet, which is 122 exactly the NAT gateway that the client wishes to communicate with 123 when managing its port mappings. 125 2. Benefit of using a PCP Anycast Address 127 The benefit of using an anycast address is simplicity and 128 reliability. In an example deployment scenario: 130 1. A network administrator installs a PCP-capable NAT. 132 2. An end user (who may be the same person) runs a PCP-enabled 133 application. This application can implement PCP with purely 134 user-level code -- no operating system support is required. 136 3. This PCP-enabled application sends its PCP request to the PCP 137 anycast address. This packet travels through the network like 138 any other, without any special support from DNS, DHCP, other 139 routers, or anything else, until it reaches the PCP-capable NAT, 140 which receives it, handles it, and sends back a reply. 142 Using the PCP anycast address, the only two things that need to be 143 deployed in the network are the two things that actually use PCP: The 144 PCP-capable NAT, and the PCP-enabled application. Nothing else in 145 the network needs to be changed or upgraded, and nothing needs to be 146 configured, including the PCP client. 148 3. Historical Objections to Anycast 150 In March 2001 a draft document entitled "Analysis of DNS Server 151 Discovery Mechanisms for IPv6" [DNSDisc] proposed using anycast to 152 discover DNS servers, a proposal that was subsequently abandoned in 153 later revisions of that draft document. 155 There are legitimate reasons why using anycast to discover DNS 156 servers is not compelling, mainly because it requires explicit 157 configuration of routing tables to direct those anycast packets to 158 the desired DNS server. However, DNS server discovery is very 159 different to NAT gateway discovery. A DNS server is something a 160 client explicitly talks to, via IP address. The DNS server may be 161 literally anywhere on the Internet. Various reasons make anycast an 162 uncompelling technique for DNS server discovery: 164 o DNS is a pure application-layer protocol, running over UDP. 166 o On an operating system without appropriate support for configuring 167 anycast addresses, a DNS server would have to use something like 168 Berkeley Packet Filter (BPF) to snoop on received packets to 169 intercept DNS requests, which is inelegant and inefficient. 171 o Without appropriate routing changes elsewhere in the network, 172 there's no reason to assume that packets sent to that anycast 173 address would even make it to the desired DNS server machine. 174 This places an addition configuration burden on the network 175 administrators, to install approprate routing table entries to 176 direct packets to the desired DNS server machine. 178 In contrast, a NAT gateway is something a client's packets stumble 179 across as they try to leave the local network and head out onto the 180 public Internet. The NAT gateway has to be on the path those packets 181 naturally take or it can't perform its NAT functions. As a result, 182 the objections to using anycast for DNS server discovery do not apply 183 to PCP: 185 o No routing changes are needed (or desired) elsewhere in the local 186 network, because the whole *point* of using anycast is that we 187 want the client's PCP request packet to take the same forwarding 188 path through the network as a TCP SYN to any other remote 189 destination address, because we want the *same* NAT gateway that 190 would have made a mapping in response to receiving an outbound TCP 191 SYN packet from the client to be the the one that makes a mapping 192 in response to receiving a PCP request packet from the client. 194 o A NAT engine is already snooping on (and rewriting) every packet 195 it forwards. As part of that snooping it could trivially look for 196 packets addressed to the PCP UDP port and process them locally 197 (just like the local processing it already does when it sees an 198 outbound TCP SYN packet). 200 4. IANA Considerations 202 IANA should allocate an IPv4 and an IPv6 well-known PCP anycast 203 address. 205 192.0.0.0/24 and 2001:0000::/23 are reserved for IETF Protocol 206 Assignments, as listed at 207 and 208 210 Suitable addresses in these ranges, such as 192.0.0.8, and a 211 corresponding suitable IPv6 address, should be allocated. 213 5. Security Considerations 215 In a network without any border gateway, NAT or firewall that is 216 aware of the PCP anycast address, outgoing PCP requests could leak 217 out onto the external Internet, possibly revealing information about 218 internal devices. 220 Using an IANA-assigned well-known PCP anycast address enables border 221 gateways to block such outgoing packets. In the default-free zone, 222 routers should be configured to drop such packets. Such 223 configuration can occur naturally via BGP messages advertising that 224 no route exists to said address. 226 Sensitive clients that do not wish to leak information about their 227 presesence can set an IP TTL on their PCP requests that limits how 228 far they can travel into the public Internet. 230 6. References 232 6.1. Normative References 234 [RFC6887] Wing, D., Cheshire, S., Boucadair, M., Penno, R., and P. 235 Selkirk, "Port Control Protocol (PCP)", RFC 6887, 236 April 2013. 238 6.2. Informative References 240 [DNSDisc] Hagino, J. and D. Thaler, "Analysis of DNS Server 241 Discovery Mechanisms for IPv6", 242 draft-ietf-ipngwg-dns-discovery-01 (work in progress), 243 November 2001. 245 [Recurs] Cheshire, S., "Recursive PCP", 246 draft-cheshire-recursive-pcp-02 (work in progress), 247 Mar 2013. 249 Author's Address 251 Stuart Cheshire 252 Apple Inc. 253 1 Infinite Loop 254 Cupertino, California 95014 255 USA 257 Phone: +1 408 974 3207 258 Email: cheshire@apple.com