Internet DRAFT - draft-grow-bounded-longest-match
draft-grow-bounded-longest-match
Network Working Group T. Hardie
Internet Draft R. White
Expiration Date: December 2003 June 2003
File Name: draft-grow-bounded-longest-match-04.txt
Bounding Longest Match Considered
draft-grow-bounded-longest-match-04.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. Internet Drafts may be updated, replaced, or obsoleted by
other documents at any time. It is not appropriate to use Internet
Drafts as reference material or to cite them other than as a "working
draft" or "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.
Abstract
Some ASes currently use length-based filters to manage the size of
the routing table they use and propagate. This draft explores an
alternative to length-based filters which allows for more automatic
configuration and which provides for better redundancy.
Rather than use a filter, this draft proposes a method of modifying
the BGP longest match algorithm by setting a bound on the prefix
lengths eligible for preference. A bound would operate on long
prefixes when covering route announcements are available; in certain
circumstances it would cause a router to prefer an aggregate over a
more specific route announcement.
Hardie, White [Page 1]
INTERNET DRAFT Bounding Longest Match Considered June 2003
1. Motivation
Modifying longest match would limit the rate of growth in the routing
table seen by many BGP speakers. The current rate of growth and the
time to convergence represent threats to the stability to the
Internet. In the short term, the IETF is considering efforts to curb
these threats while new routing paradigms that attack the fundamental
limitations of path vector protocols are developed and deployed.
A number of the practical efforts to limit the rate of growth of the
routing table have focused on filter policies, arguing that
aggressive filtering will return the Internet to a state in which
provider aggregates are a majority of the routes in the routing
table[3]. This draft proposes an approach along those same lines,
but using a bound on the longest match algorithm rather than a filter
policy. The authors believe that this approach can produce a similar
(though not identical) effect while retaining full reachability and
allowing multi-homing non-transit networks to achieve the main goals
which have motivated their becoming independent ASes.
2. Proposed Enhancements
Two enhancements are proposed by this draft: a new community, and a
new way of handling overlapping prefixes received from an external
peer.
As each prefix is received by a BGP speaker from an external peer, it
would be evaluated in the light of other prefixes already received.
If two prefixes overlap in space (such as 192.168.0.0/16 and
192.168.1.0/24), the longer prefix would be marked with the NO_EXPORT
community, and the local preference set to a very high number so that
it would always win in any best path computations within the
autonomous system. The longer prefix may also be marked with a new
community, NO_INSTALL.
2.1. The NO_INSTALL Community
An optional optimization to bounding longer prefixes by marking them
with a high Local Preference and the NO_EXPORT community is to also
mark them with a new, non-trasitive, optional community, NO_INSTALL.
The effect of this community would be for any BGP speaker receiving a
prefix with this community set to treat the prefix normally in the
BGP bestpath computation, and to forward bestpaths marked as
NO_INSTALL to iBGP peers, but to simply not install such prefixes in
the local routing table.
Hardie, White [Page 2]
INTERNET DRAFT Bounding Longest Match Considered June 2003
This would result in saving some small amount of memory for each
prefix not installed in the RIB, and the local forwarding tables
built from the RIB. If there are enough prefixes thus marked, the
memory and computation savings could be significant. BGP speakers
which receive a prefix marked with NO_INSTALL, and do not understand
this community, may simply ignore the community.
2.2. Example of Bounding the Longer Prefix
Assume the following configuration of autonomous systems:
( )
/-------( AS2 )--------\
( ) / ( ) ( R1 ) ( )
( AS1 ) ( AS4 R2)-----( AS5 )
( ) \ ( ) ( R3 ) ( )
\-------( AS3 )--------/
( )
o AS1 is advertising 192.168.1.0/24 to both AS2 and AS3.
o AS2 is advertising both 192.168.1.0/24 and 192.168.0.0/16 into
AS4
o AS3 is advertising 192.168.1.0/24 into AS4
When R1 receives both the 192.168.1.0/24 and the 192.168.0.0/16 pre-
fixes, it will mark the 192.168.1.0/24 as NO_EXPORT, and set the
local preference to a high value, as described in the section Setting
the Local Preference, below, and will then propogate this through
AS4.
R3 will receive the longer prefix from AS3, and the iBGP prefix with
the high local preference with NO_EXPORT set. Given it does not see
the overlapping prefix, it will compare the default (lower) local
preference of the externally learned route with the higher local
preference set by the AS2/AS4 border router, and will not advertise
the 192.168.1.0/24 prefix into AS4 at all.
R3 border router may also, on detecting the overlap, mark the longer
prefix with the NO_INSTALL community.
If the link between AS1 and AS2 fails, the longer length prefix will
be withdrawn from AS2, and thus the peering point between AS2 and AS4
will no longer have an overlapping set of prefixes. Within AS4, the
border router which peers with AS2 will cease advertising the
Hardie, White [Page 3]
INTERNET DRAFT Bounding Longest Match Considered June 2003
192.168.1.0/24 prefix, which allows the AS3/AS4 border router to
begin advertising it into AS4, and through AS4 into AS5, restoring
connectivity to AS1.
2.3. Setting the Local Preference
Since there could be multiple points at which an autonomous system
may receive the same pair of overlapping prefixes, there must be some
way to ensure that one of the longer prefixes wins in the BGP deci-
sion algorithm consistently. In practice, this means that each BGP
speaker which receives an overlapping set of routes should set the
local preference on the set of longer prefixes so there won't be two
longer prefixes with matching local preferences.
The easiest way to ensure this within an autonomous system is to set
the local preference for longer prefixes based on some unique number
assigned to each BGP speaker. Given the router ID and the local
preference are both 32 bit numbers, an ideal solution appears to be
to simply set the local preference to the router ID of the BGP
speaker. The primary problem with this is that in some cases, the
router ID of the device may be lower than some standard Local Prefer-
ence, perhaps even lower than a standard Local Prference used by
default throughout a network.
To alleviate this problem, the local preference of longer prefixes
which overlap with shorter prefixes should be set to the router ID of
the BGP speaker, and then the high order bit of the Local Preference
should be set, so the setting will be guaranteed to be at least above
64,000.
2.4. Implications for Load Sharing
Since the goal of this proposal is to reduce the number of paths
stored within local tables, and to reduce the amount of information
passed through to neighboring autonomous systems, the implementation
of this draft as described above would have a negative impact on the
ability to load share between multiple paths to the same destination.
Hardie, White [Page 4]
INTERNET DRAFT Bounding Longest Match Considered June 2003
3. An Alternative Implementation Using ADD PATH
An implementation which supports [ADD-PATH] could optionally use this
capability to block the overlapping prefixes into neighboring auto-
nomous systems, and preserve local load sharing.
o Any router receiving a pair of overlapped routes from its exter-
nal peers would mark the longer prefix with the NO_EXPORT com-
munity, and propogate the overlapped prefix using the technique
described in [ADD-PATH].
o Any router receiving a pair of overlapped routes, with the
longer prefix learned from an external peer, and the shorter
prefix learned from an internal peer, would mark the longer pre-
fix with the NO_EXPORT community, and propogate the prefix nor-
mally to its internal peers.
4. Benefits and Risks
The benefits and risks associated with this proposal are discussed in
the sections below.
4.1. Advantages to the Service Provider
AS4, in each of the situations, reduces the number of prefixes car-
ried through the autonomous system by the number of longer prefixes
that overlap with aggregates of those prefixes. While one copy of the
prefix continues to be carried through the autonomous system, this
entry can be marked with the optional NO_INSTALL community, so it is
not placed in the forwarding table, nor is it propogated outside the
autonomous system.
AS5 receives one prefix instead of two (or possibly more).
4.2. Advantages to the Customer
In this case, the customer is respresented as AS1. The customer will
continue to receive some amount of traffic over both peering ses-
sions, and dual homing through two Service Providers is still effec-
tive. If the customer's primary link fails, the alternate link
through AS3 will take over receving all inbound traffic automati-
cally. With most other schemes presented to this point, the customer
loses all impact of dual-homing into the Internet, unless both con-
nections are through one Service Provider.
Hardie, White [Page 5]
INTERNET DRAFT Bounding Longest Match Considered June 2003
4.3. Advantages to the Internet
Beyond the second AS hop, aggregation is preserved in all cases.
While this would not reduce the backbone routing table by the
dramatic amounts that other methods might, the advantages to the com-
munity are great, and at greatly reduced risk to customers.
4.4. Implications for Router processing
This proposal clearly adds to the work which needs to be done during
overall BGP processing. Because a check needs to be done for both
covered and covering routes, some part of this work is required for
routes of lengths on either side of the bound. Should this become
common, however, the rate of growth in the number of routes should be
smaller and a balance should be struck between the extra processing
per route and the number of routes.
4.5. Implications for Traffic engineering
The implementation of a bound risks magnifying or removing the effect
of certain widely deployed traffic engineering methods. If, for
example, an AS chose to prepend its own route to an announcement in
order to alter the preference for that route, a BGP neighbor using a
bounded longest match might now see that route as eligible for dis-
card in favor of an aggregate. While it is fairly easy to code
around that particular problem, to avoid this class of problems it
might be preferable to allow this to apply to specific AS Sets as
well as to all BGP neighbors.
4.6. Implications for Propagation delay and increased convergence time.
If the route to the AS providing the route to the aggregate should be
lost, the more-specific must propagate into the ASes which had form-
erly heard only the aggregate. This increases convergence time and
may create situations in which reachability is temporarily comprom-
ised. Unlike the filter case, however, normal BGP behavior should
restore reachability without changes to the router configuration.
There is a also a risk that during a pathological event the increased
processing required by this change will degrade propagation times
during those events. This depends on both the speed of specific
implementations and the character of the topology.
Hardie, White [Page 6]
INTERNET DRAFT Bounding Longest Match Considered June 2003
5. Security Considerations
This document presumes that the implementation of bounded longest
match is a knob inside a router config. Since the use of the knob
affects route announcements not originating within the router's AS or
its direct neighbors, the new behavior may result in surprises to the
announcing AS. It is possible that this behavior might be considered
a denial of service or mistaken for a denial of service by systems
designed to detect black-holing on behalf of the origin AS.
6. Acknowledgements
Cengiz Alaentinoglu, Alvaro Retana, Daniel Walton, Danny McPherson,
and Barry Greene gave valuable comments on this draft. A number of
colleagues also gave the author valuable comments on the white board
markings that gave rise to this paper; among them are Lane Patterson,
Ian Cooper, Gerd Besch, Bill Norton, Diarmuid Flynn, and Sean
Donelan.
7. References
[1] Huston, Geoff. http://www.telstra.net/ops/bgp/index.html
[2] Ahuja, Abha. http://www.merit.edu/~ahuja/ptomaine-bof/ahuja-ietf-
ptomaine/index.htm
[3] Bush, Randy. Plenary, IETF 51. Eventually at:
http://www.ietf.org/proceedings/01aug/
[ADD-PATH]
Walton, D, et al, "Advertisement of Multiple Paths in BGP," draft-
walton-bgp-add-paths-00.txt
Hardie, White [Page 7]
INTERNET DRAFT Bounding Longest Match Considered June 2003
8. Authors' Addresses
Ted Hardie
Ted.Hardie@nominum.com
Russ White
Cisco Systems, Inc.
7025 Kit Creek Rd.
Research Triangle Park, NC 27709
EMail: riw@cisco.com
Hardie, White [Page 8]
