IPv6 Working Group D. Thaler
INTERNET-DRAFT M. Talwar
October 25, 2003 Microsoft
Expires April 2004 C. Patel
All Play, No Work
Bridge-like Neighbor Discovery Proxies (ND Proxy)
<draft-thaler-ipv6-ndproxy-01.txt>
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.
Expires December 2003 [Page 1]
=0C
Draft ND Proxy October 2003
Abstract
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
In the IPv4 Internet today, it is common for Network Address
Translators (NATs) [NAT] to be used to easily connect one or more
leaf links to an existing network without requiring any
coordination with the network service provider. Since NATs modify
IP addresses in packets, they are problematic for many IP
applications. As a result, it is desirable to address the problem
(for both IPv4 and IPv6) without the need for NATs.
Another common solution is IEEE 802 bridging, as specified in
[BRIDGE]. It is expected that whenever possible links will be
bridged at the link layer using classic bridge technology [BRIDGE]
as opposed to using the mechanisms herein. However, classic
bridging at the data-link layer has the following limitations
(among others):
o It requires the ports to support promiscuous mode.
o It requires all ports to support the same type of link-layer
addressing (in particular, IEEE 802 addressing).
As a result, two common scenarios, described below, are not
solved, and it is these two scenarios we specifically target in
this document. While the mechanism described herein may apply to
other scenarios as well, we will concentrate our discussion on
these two scenarios.
1.1. SCENARIO 1: Wireless upstream
The following figure illustrates a likely example:
Expires December 2003 [Page 2]
=0C
Draft ND Proxy October 2003
| +-------+ +--------+
local |Ethernet | | Wireless | Access |
+---------+ A +-))) (((-+ +--> rest of network
hosts | | | link | Point |
| +-------+ +--------+
In this scenario, the access point has assigned an IPv4 and/or an
IPv6 subnet prefix to the wireless link, and uses link-layer
encryption so that wireless clients may not see each other's data.
Classic bridging requires the bridge (node A in the above diagram)
to be in promiscuous mode. In this wireless scenario, A cannot
put its wireless interface into promiscuous mode, since one
wireless node cannot see traffic to/from other wireless nodes.
This document describes a solution for both IPv4 and IPv6 which
does not involve NAT or require any change to the access point or
router.
IPv4 ARP proxying has been used for some years to solve this
problem, but the behavior has not been described in a
specification. In this document, we describe how this may be
implemented, and also enable equivalent functionality for IPv6 to
remove this incentive to deploy NATs in IPv6.
1.2. SCENARIO 2: PPP upstream
The following figure illustrates another likely example:
| +-------+ +--------+
local |Ethernet | | PPP link | |
+---------+ A +-----------+ Router +--> rest of network
hosts | | | | |
| +-------+ +--------+
In this scenario, the router believes that the other end of the
PPP link (node A) has a single IPv4 address, as negotiated by PPP.
For IPv6, it has assigned a /64 to the link and advertises it in
an IPv6 Router Advertisement.
Classic bridging does not support non-802 media, and hence IPv4
connectivity is solved by making the proxy (node A in the above
diagram) be a NAT. This document does not specify any other IPv4
solution for this scenario. However, this document does specify a
solution for IPv6 which does not involve NAT or require any change
to the router.
Expires December 2003 [Page 3]
=0C
Draft ND Proxy October 2003
2. Terminology
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL
NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL"
in this document are to be interpreted as described in BCP 14, RFC
2119 [KEYWORDS].
The term "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 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.
3. 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.
Note, this document does not support bridging of non-802
media for IPv4.
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.
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.
o Prevent loops.
o Also work in the absense of any routers.
o Support secure IPv6 neighbor discovery. This is discussed in
the Security Considerations section.
Expires December 2003 [Page 4]
=0C
Draft ND Proxy October 2003
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.
o Allow dynamic addition/removal of a proxy without adversely
disrupting the network.
o The proxy behavior should not break any existing classic
bridges in use on a network segment.
3.1. Non-requirements
The following items are not considered requirements, as they are
not met by classic bridges:
o Show up as a hop 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:
Expires December 2003 [Page 5]
=0C
Draft ND Proxy October 2003
o Support network-layer protocols other than IPv4 and IPv6. We
do not preclude such support, but it is not specified in this
document.
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 may be desirable, no solution is
currently known which does not have undesirable side effects
outside the subnet. As a result, this scenario is outside
the scope of this document.
4. Bridge-Like Proxy Behavior
Network layer support for proxying between multiple interfaces
SHOULD be used only when classic bridging is not possible.
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.
As with all other interfaces, IPv4 and IPv6 maintain a neighbor
cache (aka "ARP cache") for each proxy interface, which will be
used as described below.
4.1. Receiving Packets
When any IP or ARP packet is received on a proxy interface, it
must be parsed to see whether it is known to be of a type that
negotiates link-layer addresses. This document covers the
following types: ARP, IPv6 Neighbor Discovery, IPv6 Router
Discovery, IPv6 Redirects, and 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.
Expires December 2003 [Page 6]
=0C
Draft ND Proxy October 2003
When any other IP broadcast or multicast packet (other than to the
IPv6 Link-scoped STP Multicast Group) is received on a proxy
interface, in addition to any normal IP behavior such as 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.
The link layer header, and the link-layer address within the
payload for each forwarded packet will be modified as follows:
1) The source address will be the address of the outgoing
interface.
2) The destination address will be the address in the neighbor
entry corresponding to the destination IP address.
3) The link-layer address within the payload is substituted with
the address of the outgoing interface.
4.1.1. Proxying Packets With Link-Layer Addresses
Once it is determined that the packet is either
multicast/broadcast or else is not locally destined (if unicast),
the special types enumerated above (ARP, etc.) that carry link-
layer addresses are handled by generating a proxy packet that
contains the proxy's link-layer address on the outgoing interface
instead. Section 7. "Guidelines to proxy developers" describes the
scenarios in which the link-layer address substitution in the
Expires December 2003 [Page 7]
=0C
Draft ND Proxy October 2003
payload should be performed.
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. Note
that any change to the length of a proxied packet, such as when
the link-layer address length changes, will require corresponding
changes to fields in the IP header, namely the IPv4 Total Length
and Header Checksum fields, or the IPv6 Payload Length field.
4.1.2. IPv4 ARP Proxying
When any IPv4 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, or DHCPv4.
4.1.2.1. ARP REQUEST Packets
If the received packet is an ARP REQUEST, the request is processed
locally but no ARP REPLY is generated immediately. Instead, the
ARP REQUEST is proxied (as described above) and the ARP REPLY 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 ARP REQUEST based on its own cache.
4.1.2.2. ARP REPLY Packets
If the received packet is an ARP REPLY, the neighbor cache on the
receiving interface is first updated as if the ARP REPLY were
locally destined, and then the ARP REPLY is proxied as described
above.
4.1.2.3. DHCPv4 Packets
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 since the response will be broadcast rather than unicast
to that hardware address. The hardware address is thus used only
as a unique identifier and hence need not be a link-layer address
Expires December 2003 [Page 8]
=0C
Draft ND Proxy October 2003
on the same segment.
4.1.3. IPv6 ND Proxying
When any IPv6 packet is received on a proxy interface, it must be
parsed to see whether it is known to be one of the following
types: IPv6 Neighbor Discovery, IPv6 Router Discovery, or IPv6
Redirects.
4.1.3.1. ICMPv6 Neighbor Solicitations
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 as
described above 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.
4.1.3.2. ICMPv6 Neighbor Advertisements
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 NA is proxied as
described above.
4.1.3.3. ICMPv6 Router Advertisements
To support scenario 2, if the MTU of the upstream segment is
smaller than the MTU of the downstream segment then IPv6 Router
Advertisements (RAs) must also be proxied as follows. If the RA
contains an MTU Option, the RA is forwarded unmodified.
Otherwise, the proxy adds an MTU Option with a value of minimum of
the link MTUs of the proxy interfaces, and then proxies it as
described above.
Note that packet loss will occur in the scenario where the proxy
has links with different MTU, and the router sets the MTU option
with size greater than the minimum of the link MTUs of the proxy.
Expires December 2003 [Page 9]
=0C
Draft ND Proxy October 2003
4.1.3.4. ICMPv6 Redirects
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.
4.2. Sending Packets
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.
5. 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
matches the on-link subnet prefix, and a destination cache entry
Expires December 2003 [Page 10]
=0C
Draft ND Proxy October 2003
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. 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.
6. Loop Prevention
Loop prevention can be done done by having the proxy implement the
Spanning Tree Algorithm and Protocol as defined in [BRIDGE] on all
proxy interfaces. Loop prevention is OPTIONAL, and is useful only
Expires December 2003 [Page 11]
=0C
Draft ND Proxy October 2003
if the proxy can be deployed in an environment where physical
loops are possible. For example, in a cell phone which proxies
between a PPP dialup link and a local Ethernet interface, it is
typically safe to assume that physical loops are not possible and
hence there is no need to support the Spanning Tree Protocol
(STP).
If loop prevention is implemented, it is done as follows. IEEE
802 interfaces use the protocol exactly as specified in [BRIDGE].
Operation of the Spanning Tree Protocol (STP) over other types of
link layers is done by encapsulating the STP frame in an IPv6
header as follows. The Next Header field is set to [TBA by IANA],
indicating that an STP header follows. The Destination Address
field is set to the Link-scoped STP Multicast Group [TBA by IANA].
All proxies operating on non-IEEE 802 media join this group so
they will receive STP packets. STP packets are never forwarded or
proxied.
7. Guidelines to proxy developers
Proxy developers will have to accomodate protocol or protocol
options (for example, new ICMP messages) that are developed in the
future, or protocols that are not mentioned in this document (for
example, propritery protocols). This section prescribes guidelines
that can be used by proxy developers to accomodate protocols that
are not mentioned herein.
1) If the link-layer address in the payload of the protocol can
be used in the link-layer header of future messages, then the
proxy should substitute it with its own address. For example
the link-layer address in NA messages is used in the link-
layer header for future messages, and, hence, the proxy
substitutes it with its own address.
For broadcast/multicast packets, the link-layer address
substituted within the payload will be different for each
outgoing interface.
2) If the link-layer address in the payload of the protocol will
never be used in any link-layer header, then the proxy should
not substitute it with its own address.
Expires December 2003 [Page 12]
=0C
Draft ND Proxy October 2003
8. IANA Considerations
To support loop prevention over non-802 media, IANA should assign:
1) a Protocol Number for STP, and
2) an IPv6 Link-Local Scope multicast address for All-STP-
Speakers.
9. Security Considerations
Proxies are susceptible to the same kind of security issues that
plague hosts using unsecured Neighbor Discovery or ARP. Even if
these protocols are secured, the proxies may process unsecured
messages, and update the neighbor cache. Malicious nodes within
the subnet can take advantage of this property, and hijack
traffic. The threats are discussed in detail in [PSREQ].
As a result, securing Neighbor Discovery or ARP must take into
account the ability to proxy messages. This document does not
introduce any new requirements in this regard.
From=20an IPv6 perspective, RFC 2461 [ND] already defines the
ability to proxy Neighbor Advertisements. The requirements for
securing proxied messages 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.
10. Appendix A: Comparison with other approaches
It has been suggested that a simple Router Advertisement (RA)
proxy would be sufficient, where the subnet prefix in an RA is
"stolen" by the proxy and applied to a downstream link instead of
an upstream link. Other ND messages are not proxied.
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. This is because when a node on a downstream link
tries to do Neighbor Discovery, and the proxy does not send the NS
on the upstream link, it will never discover the neighbor on the
Expires December 2003 [Page 13]
=0C
Draft ND Proxy October 2003
upstream link. Similarly, if messages are not proxied during DAD,
conflicts can occur.
Second, if the proxy assumes that no nodes on the upstream link
have addresses in the prefix, such a proxy could not be safely
deployed without cooperation from the network administrator since
it introduces a requirement that the router itself not have an
address in the prefix. This rules out out use in situations where
bridges and Network Address Translators (NATs) are used today,
which is the problem this document is directly addressing.
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.
11. 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
Chirayu Patel
All Play, No Work
Bangalore, Karnataka 560038
Phone: +91-98452-88078
EMail: chirayu@chirayu.org
12. Normative References
[ARP]
D. Plummer, "An Ethernet Address Resolution Protocol", STD
37, RFC 826, November 1982.
Expires December 2003 [Page 14]
=0C
Draft ND Proxy October 2003
[BRIDGE]
T. Jeffree, editor, "Media Access Control (MAC) Bridges",
ANSI/IEEE Std 802.1D, 1998,
http://standards.ieee.org/getieee802/download/802.1D-1998.pdf.
[DHCPv4]
R. Droms, "Dynamic Host Configuration Protocol", RFC 2131,
March 1997.
[KEYWORDS]
S. Bradner, "Key words for use in RFCs to Indicate
Requirement Levels", BCP 14, RFC 2119, March, 1997.
[ND] Narten, T., Nordmark, E., and W. Simpson, "Neighbor Discovery
for IP Version 6 (IPv6)", RFC 2461, December 1998.
13. Informative References
[NAT]
Srisuresh, P., and K. Egevang, "Traditional IP Network
Address Translator (Traditional NAT)", RFC 3022, January
2001.
[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.
14. 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
Expires December 2003 [Page 15]
=0C
Draft ND Proxy October 2003
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.
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 16]