IPv6 Working Group D. Thaler INTERNET-DRAFT M. Talwar Expires December 2003 Microsoft June 9, 2003 Bridge-like Neighbor Discovery Proxies (ND Proxy) Status of this Memo This document is an Internet-Draft and is in full conformance with all provisions of Section 10 of RFC2026. Internet-Drafts are working documents of the Internet Engineering Task Force (IETF), its areas, and its working groups. Note that other groups may also distribute working documents as Internet- Drafts. Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet- Drafts as reference material or to cite them other than as "work in progress." The list of current Internet-Drafts can be accessed at http://www.ietf.org/ietf/1id-abstracts.txt The list of Internet-Draft Shadow Directories can be accessed at http://www.ietf.org/shadow.html. Copyright Notice Copyright (C) The Internet Society (2003). All Rights Reserved. Abstract Expires December 2003 [Page 1] Draft ND Proxy March 2003 Bridging multiple links into a single entity has several operational advantages. A single subnet prefix is sufficient to support multiple physical links. There is no need to allocate subnet numbers to the different networks, simplifying management. Bridging some types of media requires network-layer support, however. This document describes these cases and specifies the IP-layer support that enables bridging under these circumstances. 1. Introduction IEEE 802 bridging, specified in [BRIDGE] is in widespread usage today. However, classic bridging at the data-link layer has the following limitations (among others): o It requires the ports to support promiscuous mode. As a result, an 802.11 segment cannot be bridged except at the access point. o It requires all ports to support the same type of link-layer addressing (in particular, IEEE 802 addressing). For example, an Ethernet segment and a PPP segment cannot be bridged. Under such conditions, multiple segments can still be bridged by detecting and proxying specific messages at the network layer, and forwarding all others. In the document, we specify the behavior for such a proxy for IPv6 as well as IPv4, which is indistinguishable (as much as possible) from a classic bridge. It is expected that links will be bridged at the link layer using classic bridge technology whenever possible and a single "bridge" interface will be exposed to the IP layer, and that network layer support for proxying between multiple interfaces is used only when this is not possible. In the remainder of this document, a "proxy interface" will be used to refer to an interface (which could itself be a bridge interface) over which network layer proxying is done as defined herein. In contrast to [MLSR], in this document we make no distinction between a "link" (in the classic IPv6 sense) and a "subnet". We use the term "segment" to apply to a bridged component of the link. It has been suggested that a simple Router Advertisement (RA) proxy would be sufficient, where the subnet prefix in an RA is Expires December 2003 [Page 2] Draft ND Proxy March 2003 "stolen" by the proxy and applied to a downstream link _instead of_ an upstream link. There are many problems with this approach. First, it requires cooperation from all nodes on the upstream link. No node (including the router sending the RA) can have an address in the subnet or it will not have connectivity with nodes on the downstream link. Second, as a result, such a proxy could not be used without cooperation from the network administrator, ruling out use in situations where bridges and Network Address Translators (NATs) are used today. Instead, where a prefix is desired for use on one or more downstream links in cooperation with the network administrator, Prefix Delegation should be used instead. We now discuss requirements for the bridge-like proxy mechanism. 1.1. Requirements Bridge-like proxy behavior is designed with the following requirements in mind: o Support connecting multiple segments with a single subnet prefix. o Support media which cannot be bridged at the link-layer. o Do not require any changes to existing routers. That is, any routers on the subnet should be unaware that the subnet is being bridged. It should appear as if one host uses multiple addresses. o Provide full connectivity. For example, if there are existing nodes (such as any routers on the subnet) which have addresses in the subnet prefix, adding a bridge-like proxy must allow bridged nodes to have full connectivity with existing nodes on the subnet. If, on the other hand, neighbor discovery messages were not proxied on the segment containing a router or other node with an existing address, then confusion, duplicate addresses, and lack of connectivity could result. o Prevent loops. o Also work in the absense of any routers. Expires December 2003 [Page 3] Draft ND Proxy March 2003 o Support secure IPv6 neighbor discovery. This is discussed in the Security Considerations section. o Support both IPv6 and IPv4. o Support nodes moving between segments. For example, a node should be able to keep its address without seeing its address as a duplicate due to any cache maintained at the proxy. 1.2. Non-requirements The following items are not considered requirements, as they are not met by classic bridges: o Show up in a traceroute. o Use the shortest path between two nodes on different segments. o Be able to use all available interfaces simultaneously. Instead, bridging technology relies on disabling redundant interfaces to prevent loops. o Support differing MTUs in use on different segments. That is, all segments on a bridged link must use the smallest MTU of any segment. Note that the result of this is that in the absence of cooperation of the network administrator (who can configure routers with a smaller MTU to advertise in Router Advertisements) a bridge-like IPv6 proxy can only connect links with equal MTU, or where all routers are on segments with the smallest MTU. o Support connecting media on which Neighbor Discovery is not possible. For example, some technologies such as 6to4 use an algorithmic mapping from IPv6 address to the underlying link- layer (IPv4 in this case) address, and hence cannot support bridging arbitrary IP addresses. The following additional items are not considered requirements for this document: o Support network-layer protocols other than IPv4 and IPv6. We do not preclude such support, but it is not specified in this document. Expires December 2003 [Page 4] Draft ND Proxy March 2003 o Support Neighbor Discovery, Router Discovery, or DHCPv4 packets using encryption with an ESP header. We also note that the current methods for securing these protocols do not use an ESP header. Where encryption is required, link-layer encryption can be used on each segment. o Support Redirects for off-subnet destinations that point to a router on a different segment from the redirected host. While this scenario would be desirable, no solution is currently known which does not have undesirable side effects outside the subnet. As a result, this feature is left as an opportunity for future work. 2. Bridge-Like Proxy Behavior In general, the proxy attempts to emulate a bridge as much as possible, and augments this with IP-specific behavior for cases which could not otherwise be handled. When a proxy interface comes up, the node puts it in "all- multicast" mode so that it will receive all multicast packets. It is common for interfaces to not support full promiscuous mode (e.g., on a wireless client), but all-multicast mode is generally still supported. Loop prevention is done by having the proxy implement the Spanning Tree Algorithm and Protocol as defined in [BRIDGE] on all proxy interfaces. As with all other interfaces, IPv4 and IPv6 maintain a neighbor cache for each proxy interface, which will be used as described below. When any IP or ARP packet is received on a proxy interface, it must be parsed to see whether it is known to be one of the following types: ARP, IPv6 Neighbor Discovery, IPv6 Router Discovery, IPv6 Redirects, or DHCPv4. These packets are ones that can carry link-layer addresses, and hence must be proxied (as described below) so that packets between nodes on different segments can be received by the proxy and have the correct link- layer address type on each segment. When any other IP broadcast or multicast packet is received on a proxy interface, in addition to any normal IP behavior such as Expires December 2003 [Page 5] Draft ND Proxy March 2003 being delivered locally, it is forwarded unchanged out all other proxy interfaces on the same link. (As specified in [BRIDGE], the proxy may instead support multicast learning and filtering but this is optional.) In particular, the IPv4 TTL or IPv6 Hop Limit is not updated, and no ICMP errors are sent as a result of attempting this forwarding. When any other IP unicast packet is received on a proxy interface, if it is not locally destined then it is forwarded unchanged to the proxy interface for which the next hop address appears in the neighbor cache. Again the IPv4 TTL or IPv6 Hop Limit is not updated, and no ICMP errors are sent as a result of attempting this forwarding. To choose a proxy interface to forward to, the neighbor cache is consulted, and the interface with the neighbor entry in the "best" state is used. In order of least to most preferred, the states (per [ND]) are INCOMPLETE, STALE, DELAY, PROBE, REACHABLE. A packet is never forwarded back out the same interface on which it arrived; such a packet is instead silently dropped. Locally originated packets that are sent on a proxy interface also follow the same rules as packets received on a proxy interface. If no neighbor entry exists when a packet is to be locally originated, an interface can be chosen in any implementation- specific fashion. Once the neighbor is resolved, the actual interface will be discovered and the packet will be sent on that interface. The special types enumerated above (ARP, etc.) that carry link- layer addresses are handled, once it is determined that the packet is either multicast/broadcast or else is not locally destined (if unicast), by generating a proxy packet that contains the proxy's link-layer address instead. As with all forwarded packets, the link-layer header is also new. Any Authentication Header would also be removed, and a new one may be added as discussed below under Security Considerations. In addition, if the received packet is an ICMPv6 Neighbor Solicitation, the NS is processed locally as described in section 7.2.3 of [ND] but no NA is generated immediately. Instead the NS is proxied and the NA will be proxied when it is received. This ensures that the proxy does not interfere with hosts moving from one segment to another since it never responds to an NS based on its own cache. Expires December 2003 [Page 6] Draft ND Proxy March 2003 If the received packet is an ICMPv6 Neighbor Advertisement, the neighbor cache on the receiving interface is first updated as if the NA were locally destined, and then the Override bit is cleared in the proxied packet. For IPv4, ARP packets are similarly proxied (except that no Override bit exists to clear). If the received packet is a DHCPv4 DISCOVER or REQUEST message, then instead of changing the client's hardware address in the payload, the BROADCAST (B) flag is set in the proxied packet. This ensures that the proxy will be able to receive and proxy the response. If the received packet is an ICMPv6 Redirect message, then the proxied packet should be modified as follows. If the proxy has a valid (i.e., not INCOMPLETE) neighbor entry for the target address on the same interface as the redirected host, then the TLLA option in the proxied Redirect simply contains the link-layer address of the target as found in the proxy's neighbor entry, since the redirected host may reach the target address directly. Otherwise, if the proxy has a valid neighbor entry for the target address on some other interface, then the TLLA option in the proxied packet contains the link-layer address of the proxy on the sending interface, since the redirected host must reach the target address through the proxy. Otherwise, the proxy has no valid neighbor entry for the target address, and the proxied packet contains no TLLA option, which will cause the redirected host to perform neighbor discovery for the target address. 2.1. Example Consider the following topology, where A and B are nodes on separate segments which are connected by a bridge-like proxy P: A---|---P---|---B a p1 p2 b A and B have link-layer addresses a and b, respectively. P has link-layer addresses p1 and p2 on the two segments. We now walk through the actions that happen when A attempts to send an initial IPv6 packet to B. A first does a route lookup on the destination address B. This Expires December 2003 [Page 7] Draft ND Proxy March 2003 matches the on-link subnet prefix, and a destination cache entry is created as well as a neighbor cache entry in the INCOMPLETE state. Before the packet can be sent, A needs to resolve B's link-layer address and sends a Neighbor Solicitation (NS) to the solicited-node multicast address for B. The SLLA option in the solicitation contains A's link-layer address. P receives the solicitation (since it is receiving all link-layer multicast packets) and processes it as it would any multicast packet by forwarding it out to other segments on the link. However, before actually sending the packet, it determines if the packet being sent is one which requires proxying. Since it is an NS, it creates a neighbor entry for A on interface 1 and records its link-layer address. It also creates a neighbor entry for B (on an arbitrary proxy interface) in the INCOMPLETE state. Since the packet is multicast, P then needs to proxy the NS out all other proxy interfaces on the subnet. Before sending the packet out interface 2, it replaces the link-layer address in the SLLA option with its own link-layer address, p2. B receives this NS, processing it as usual. Hence it creates a neighbor entry for A mapping it to the link-layer address p2. It responds with a Neighbor Advertisement (NA) sent to A containing B's link-layer address b. The NA is sent using A's neighbor entry, i.e. to the link-layer address p2. The NA is received by P, which then processes it as it would any unicast packet; i.e., it forwards this out interface 1, based on the neighbor cache. However, before actually sending the packet out, it inspects it to determine if the packet being sent is one which requires proxying. Since it is an NA, it updates its neighbor entry for B to be REACHABLE and records the link-layer address b. P then replaces the link-layer address in the TLLA option with its own link-layer address on the outgoing interface, p1. It also clears the Override bit, since the NA is being proxied. The packet is then sent out interface 1. A receives this NA, processing it as usual. Hence it creates a neighbor entry for B on interface 2 in the REACHABLE state and records the link-layer address p1. Expires December 2003 [Page 8] Draft ND Proxy March 2003 3. Security Considerations Securing neighbor discovery must take into account the ability to proxy messages. This document does not introduce any new requirements in this regard, since RFC 2461 [ND] already defines the ability to proxy Neighbor Advertisements, specifying that the Override bit is always clear in a proxied advertisement. When a receiver sees that the Override bit is clear, however, it typically cannot tell whether the advertisement was for an anycast address, or whether the advertisement was proxied, or both. As a result, secure neighbor discovery must take this into account. The threats are discussed in detail in [PSREQ]. The requirements for securing proxied Neighbor Advertisements are similar to those for securing Router Advertisements, since the receiver must verify that the advertisement came from a valid router/proxy, rather than from the owner of a specific address. 4. Authors' Addresses Dave Thaler Microsoft Corporation One Microsoft Way Redmond, WA 98052-6399 Phone: +1 425 703 8835 EMail: dthaler@microsoft.com Mohit Talwar Microsoft Corporation One Microsoft Way Redmond, WA 98052-6399 Phone: +1 425 705 3131 EMail: mohitt@microsoft.com 5. Normative References [ARP] D. Plummer, "An Ethernet Address Resolution Protocol", STD 37, RFC 826, November 1982. [BRIDGE] T. Jeffree, editor, "Media Access Control (MAC) Bridges", ANSI/IEEE Std 802.1D, 1998, Expires December 2003 [Page 9] Draft ND Proxy March 2003 http://standards.ieee.org/getieee802/download/802.1D-1998.pdf. [DHCPv4] R. Droms, "Dynamic Host Configuration Protocol", RFC 2131, March 1997. [ND] Narten, T., Nordmark, E., and W. Simpson, "Neighbor Discovery for IP Version 6 (IPv6)", RFC 2461, December 1998. 6. Informative References [MLSR] Thaler, D., and C. Huitema, "Multi-link Subnet Support in IPv6", Work in progress, draft-ietf-ipv6-multilink- subnets-00.txt, June 2002. [PSREQ] Nikander, P., Kempf, J., and E. Nordmark, "IPv6 Neighbor Discovery trust models and threats", Work in progress, draft- ietf-send-psreq-03.txt, April 2003. 7. Full Copyright Statement Copyright (C) The Internet Society (2003). All Rights Reserved. This document and translations of it may be copied and furnished to others, and derivative works that comment on or otherwise explain it or assist in its implmentation may be prepared, copied, published and distributed, in whole or in part, without restriction of any kind, provided that the above copyright notice and this paragraph are included on all such copies and derivative works. However, this document itself may not be modified in any way, such as by removing the copyright notice or references to the Internet Society or other Internet organizations, except as needed for the purpose of developing Internet standards in which case the procedures for copyrights defined in the Internet Standards process must be followed, or as required to translate it into languages other than English. Expires December 2003 [Page 10] Draft ND Proxy March 2003 The limited permissions granted above are perpetual and will not be revoked by the Internet Society or its successors or assigns. This document and the information contained herein is provided on an "AS IS" basis and THE INTERNET SOCIETY AND THE INTERNET ENGINEERING TASK FORCE DISCLAIMS ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Expires December 2003 [Page 11]