Using 127-bit IPv6 Prefixes on Inter-Router Links
draft-kohno-ipv6-prefixlen-p2p-02.txt

Abstract

On inter-router point-to-point links, it is useful for security and other reasons, to use 127-bit IPv6 prefixes. Such a practice parallels the use of 31-bit prefixes in IPv4 [RFC3021] (Retana, A., “Using 31-Bit Prefixes on IPv4 Point-to-Point Links,” December 2000.). This document outlines some of these reasons and specifies that 127-bit IPv6 prefix lengths must be supported on such links.

Status of this Memo

This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.

Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts is at http://datatracker.ietf.org/drafts/current/.

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.”

This Internet-Draft will expire on January 30, 2011.

Copyright Notice

Copyright (c) 2010 IETF Trust and the persons identified as the document authors. All rights reserved.

This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (http://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License.

Table of Contents

1.  Conventions Used In This Document
2.  Introduction
3.  Scope Of This Memo
4.  Problems identified with 127-bit prefix lengths in the past
5.  Reasons for using longer prefixes
    5.1.  Ping-pong issue
    5.2.  Neighbor Cache Exhaustion issue
    5.3.  Other reasons
6.  Recommendations
7.  Security Considerations
8.  IANA Considerations
9.  Contributors
10.  Acknowledgments
11.  References
    11.1.  Normative References
    11.2.  Informative References
§  Authors' Addresses

1.  Conventions Used In This Document

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 RFC 2119 (Bradner, S., “Key words for use in RFCs to Indicate Requirement Levels,” March 1997.) [RFC2119].

2.  Introduction

RFC 4291 (Hinden, R. and S. Deering, “IP Version 6 Addressing Architecture,” February 2006.) [RFC4291] specifies that interface IDs for all unicast address, except those that start with the binary value 000, are required to be 64 bits long and to be constructed in Modified EUI-64 format. In addition, it defines the Subnet-Router anycast address, which is intended to be used for applications where a node needs to communicate with any one of the set of routers on a link.

Some operators have been using 127-bit prefixes, but this has been discouraged due to conflicts with Subnet-Router anycast [RFC3627] (Savola, P., “Use of /127 Prefix Length Between Routers Considered Harmful,” September 2003.). However, using 64-bit prefixes creates security issues which are particularly problematic on inter-router links, and there are other valid reasons to use prefixes longer than 64 bits, in particular /127 (see Section 5 (Reasons for using longer prefixes)).

This document provides rationale for using 127-bit prefix lengths, reevaluates the reasons why doing so was considered harmful, and specifies that /127 prefixes MUST be supported on inter-router links configured for use as point-to-point links.

3.  Scope Of This Memo

This document is applicable to cases where operators assign specific addresses on inter-router point-to-point links and do not rely on link-local addresses. Many operators assign specific addresses for purposes of network monitoring, reverse DNS resolution for traceroute and other management tools, EBGP peering sessions, and so on.

For the purposes of this document, an inter-router point-to-point link is a link to which only two routers and no hosts are attached. This may include Ethernet links which are configured to be point-to-point. In such cases, there is no need to support Neighbor Discovery for address resolution, and other general scenarios like the use of stateless addresss autoconfiguration are not relevant.

Links between a router and a host, or links to which both routers and hosts are attached, are out of scope of this document.

4.  Problems identified with 127-bit prefix lengths in the past

[RFC3627] (Savola, P., “Use of /127 Prefix Length Between Routers Considered Harmful,” September 2003.) discourages the use of 127-bit prefix lengths due to conflicts with the Subnet-Router anycast addresses, while stating that the utility of Subnet-Router Anycast for point-to-point links is questionable.

[RFC5375] (Van de Velde, G., Popoviciu, C., Chown, T., Bonness, O., and C. Hahn, “IPv6 Unicast Address Assignment Considerations,” December 2008.) also says the usage of 127-bit prefix lengths is not valid and should be strongly discouraged, but the stated reason for doing this is to be in compliance with [RFC3627] (Savola, P., “Use of /127 Prefix Length Between Routers Considered Harmful,” September 2003.).

Though the analysis in the RFCs are correct, it contradicts the reality as to the /127 discussion. Given that Subnet-Router Anycast is not currently used, the RFCs could have been to recommend that Subnet-Router Anycast be disabled on prefixes that are 127 bits long.

5.  Reasons for using longer prefixes

There are various reasons for network operators to use IPv6 prefix length greater than 64, particularly 127, for inter-router links.

5.1.  Ping-pong issue

As described in [PINGPONG] (Hagino, H. and T. Jimmei, “Avoiding ping-pong packets on point-to-point links,” .), a forwarding loop may occur on a point-to-point link with a prefix length shorter than 127. This does not affect interfaces that perform Neighbor Discovery, but some point-to-point links, which uses medium such as SONET, do not use Neighbor Discovery. As a consequence, configuring any prefix length shorter than 127 bits on these links can create an attack vector in the network.

The pingpong issue happens in case of IPv4 as well. But due to the scarcity of IPv4 address space, the current practice is to assgin long prefix lengths such as /30 or /31 [RFC3021] (Retana, A., “Using 31-Bit Prefixes on IPv4 Point-to-Point Links,” December 2000.) on point-to-point links, thus the problem did not come to the fore.

The latest ICMPv6 specification [RFC4443] (Conta, A., Deering, S., and M. Gupta, “Internet Control Message Protocol (ICMPv6) for the Internet Protocol Version 6 (IPv6) Specification,” March 2006.) provides a solution to this issue by specifying that a router receiving a packet on a point-to-point link sent to an address on the link itself must not forward the packet back onto the same link, instead generating an ICMPv6 Destination Unreachable (Code 3) message. However, checking every packets for this condition could have some drawbacks.

5.2.  Neighbor Cache Exhaustion issue

As described in Section 4.3.2 of [RFC3756] (Nikander, P., Kempf, J., and E. Nordmark, “IPv6 Neighbor Discovery (ND) Trust Models and Threats,” May 2004.), the use of a 64-bit prefix length on an inter-router link that uses Neighbor Discovery (e.g., Ethernet) potentially allows for denial-of-service attacks on the routers on the link.

Consider an Ethernet link between two routers A and B to which a /64 subnet has been assigned. A packet sent to any address on the /64 (except the addresses of A and B) will cause the router attempting to forward it to create an new cache entry in state INCOMPLETE, send a Neighbor Solicitation message to be sent on the link, start a retransmit timer, and so on [RFC4861] (Narten, T., Nordmark, E., Simpson, W., and H. Soliman, “Neighbor Discovery for IP version 6 (IPv6),” September 2007.).

By sending a continuous stream of packets to a large number of the 2^64 - 3 addresses on the link that are not assigned to the routers (one for each router and one for Subnet-Router Anycast), an attacker can cause the routers to create a large number of neighbor cache entries and send a large number of Neighbor Solicitation packets which will never receive replies, possibly consuming a disproportionate amount of memory and processing resources. Sending the packets to one of the 2^24 addresses on the link that have the same Solicited-Node multicast address as of one of the routers also causes the other router to spend disproportionate amounts of processing time discarding useless Neighbor Solicitation messages.

Careful implementation and rate-limiting can limit the impact of such an attack, but are unlikely to neutralize it completely. Rate-limiting neighbor solicitation messages will reduce CPU usage, and following the garbage-collection recommendations in [RFC4861] (Narten, T., Nordmark, E., Simpson, W., and H. Soliman, “Neighbor Discovery for IP version 6 (IPv6),” September 2007.) will maintain reachability, but if the link is down and neighbor cache entries have expired while the attack is ongoing, legitimate traffic (for example, BGP sessions) over the link might never be re-established because the routers cannot resolve each others' IPv6 addresses to MAC addresses.

This attack is not specific to point-to-point links, but is particularly harmful in the case of point-to-point backbone links, which may carry large amounts of traffic to many destinations over long distances.

5.3.  Other reasons

Though address space conservation considerations are less important for IPv6 than they are in IPv4, it may still be desirable to use the smallest possible prefix to number links (and thus use, for example, /127 for point-to-point links). For example, a large end-site that is assigned a /48 of IPv6 space may not want to reserve a full /64 for every point-to-point link to avoid renumbering in the future.

6.  Recommendations

In light of the above reasons, this document specifies that 127-bit prefix lengths MUST be supported on inter-router point-to-point links, and proposes that such links MAY be assigned 127-bit prefix lengths. If such a prefix is assigned to a link, Subnet-Router Anycast MUST be disabled for the prefix.

The [RFC4291] (Hinden, R. and S. Deering, “IP Version 6 Addressing Architecture,” February 2006.)is to be revised to allow longer prefix than /64, and state that Subnet-router anycast address MUST be disabled if the prefix length of the link is /127.

7.  Security Considerations

This draft addresses, among other things, various security issues.

8.  IANA Considerations

None.

9.  Contributors

Chris Morrow, morrowc@google.com

Pekka Savola, pekkas@netcore.fi

Remi Despres, remi.despres@free.fr

Thomas Narten, narten@us.ibm.com

10.  Acknowledgments

We'd like to thank Ron Bonica, Pramod Srinivasan, Olivier Vautrin, Tomoya Yoshida, Warren Kumari and Tatsuya Jinmei for their helpful inputs.

11.  References

11.1. Normative References

11.2. Informative References

Authors' Addresses