GROW Working Group                                             R. Raszuk
Internet-Draft                                                   NTT MCL
Intended status: Informational                                     A. Lo
Expires: March 18, November 4, 2012                                            Cisco                                         Arista
                                                                L. Zhang
                                                                    UCLA
                                                                   X. Xu
                                                                  Huawei
                                                      September 15, 2011
                                                             May 3, 2012

                   Simple Virtual Aggregation (S-VA)
                    draft-ietf-grow-simple-va-04.txt
                    draft-ietf-grow-simple-va-05.txt

Abstract

   The continued growth in the Default Free Routing Table (DFRT)
   stresses the global routing system in a number of ways.  One of the
   most costly stresses is FIB size: ISPs often must upgrade router
   hardware simply because the FIB has run out of space, and router
   vendors must design routers that have adequate FIB.

   FIB suppression is an approach to relieving stress on the FIB by NOT
   loading selected RIB entries into the FIB.  Simple Virtual
   Aggregation (S-VA) is a simple form of Virtual Aggregation (VA) that
   allows any and all edge routers to shrink their RIB and FIB
   requirements substantially and therefore increase their useful
   lifetime.

   S-VA does not increase FIB requirements for core routers.  S-VA is
   extremely easy to configure considerably more so than the various
   tricks done today to extend the life of edge routers.  S-VA can be
   deployed autonomously by an ISP (cooperation between ISPs is not
   required), and can co-exist with legacy routers in the ISP.

Status of this Memo

   This Internet-Draft is submitted to IETF 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 March 18, November 4, 2012.

Copyright Notice

   Copyright (c) 2011 2012 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.  Introduction  . . . . . . . . . . . . . . . . . . . . . . . . . 4
     1.1.  Scope of this Document  . . . . . . . . . . . . . . . . . . 5
     1.2.  Requirements notation . . . . . . . . . . . . . . . . . . . 5
     1.3.  Terminology . . . . . . . . . . . . . . . . . . . . . . . . 5
   2.  Operation of S-VA . . . . . . . . . . . . . . . . . . . . . . . 6
   3.  Deployment considerations . . . . . . . . . . . . . . . . . . . 7
   4.  IANA Considerations . . . . . . . . . . . . . . . . . . . . . . 8
   5.  Security Considerations . . . . . . . . . . . . . . . . . . . . 8
   6.  Acknowledgements  . . . . . . . . . . . . . . . . . . . . . . . 8
   7.  References  . . . . . . . . . . . . . . . . . . . . . . . . . . 9
     7.1.  Normative References  . . . . . . . . . . . . . . . . . . . 9
     7.2.  Informative References  . . . . . . . . . . . . . . . . . . 9
   Authors' Addresses  . . . . . . . . . . . . . . . . . . . . . . . . 9

1.  Introduction

   ISPs today manage constant DFRT growth in a number of ways.  One way,
   of course, is for ISPs to upgrade their router hardware before DFRT
   growth outstrips the size of the FIB.  This is too expensive for many
   ISPs.  They would prefer to extend the lifetime of routers whose FIBs
   can no longer hold the full DFRT.

   A common approach taken by lower-tier ISPs is to default route to
   their providers.  Routes to customers and peer ISPs are maintained,
   but everything else defaults to the provider.  This approach has
   several disadvantages.  First, packets to Internet destinations may
   take longer-than-necessary AS paths.

   This problem can be mitigated through careful configuration of
   partial defaults, but this can require substantial configuration
   overhead.  A second problem with defaulting to providers is that the
   ISP is no longer able to provide the full DFRT to its customers.
   Finally, provider defaults prevents the ISP from being able to detect
   martian packets.  As a result, the ISP transmits packets that could
   otherwise have been dropped over its expensive provider links.

   An alternative is for the ISP to maintain full routes in its core
   routers, but to filter routes from edge routers that do not require a
   full DFRT.  These edge routers can then default route to the core or
   exit routers.  This is often possible with edge routers that
   interface to customer networks.  The problem with this approach is
   that it cannot be used for all edge routers.  For instance, it cannot
   be used for routers that connect to transits.  It should also not be
   used for routers that connect to customers which wish to receive the
   full DFRT.

   This draft describes a very simple technique, called Simple Virtual
   Aggregation (S-VA), that allows any and all edge routers to have
   substantially reduced FIB requirements even while still advertising
   and receiving the full DFRT over BGP.  The basic idea is as follows.
   Core routers in the ISP maintain the full DFRT in the FIB and RIB.
   Edge routers maintain the full DFRT in the BGP protocol RIB, but
   suppress certain routes from being installed in RIB and FIB tables.
   Edge routers install a default route to core routers, to ABRs which
   are installed on the POP to core boundary or to the ASBR routers.

   S-VA requires no changes to BGP and no changes to any choice of
   forwarding mechanisms in routers.  Configuration is extremely simple:
   S-VA must be enabled on the edge router which needs to save it's RIB
   and FIB space.  In the same time operator must inject into his intra-
   domain routing a new prefix further called virtual aggregate (VA-
   prefix) which will be used as the aggregate forwarding reference by
   the edge routers performing S-VA.  Everything else is automatic.
   ISPs can deploy FIB suppression autonomously and with no coordination
   with neighbor ASes.

1.1.  Scope of this Document

   The scope of this document is limited to Intra-domain S-VA operation.
   In other words, the case where a single ISP autonomously operates
   S-VA internally without any coordination with neighboring ISPs.

   Note that this document assumes that the S-VA "domain" (i.e. the unit
   of autonomy) is the AS (that is, different ASes run S-VA
   independently and without coordination).  For the remainder of this
   document, the terms ISP, AS, and domain are used interchangeably.

   This document applies equally to IPv4 and IPv6 both unicast and
   multicast address families.

   S-VA may operate with a mix of upgraded routers and legacy routers.
   There are no topological restrictions placed on the mix of routers.
   S-VA functionality is local to the router on which it is enabled and
   routing correctness is guaranteed.

   Note that S-VA is a greatly simplified variant of "full VA"
   [I-D.ietf-grow-va].  With full VA, all routers (core or otherwise)
   can have reduced FIBs.  However, full VA requires substantial new
   configuration and operational complexity compared to S-VA.  Full VA
   also requires the use of MPLS LSPs between all routers.  Note that
   S-VA was formerly specified in [I-D.ietf-grow-va].  It has been moved
   to this separate draft to simplify its understanding.

1.2.  Requirements notation

   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 [RFC2119].

1.3.  Terminology

   RIB/FIB-Installing Router (FIR):  An router that does not suppress
      any routes, and advertises itself as a default route for 0/0.
      Typically a core router, POP to core boundary router or an ASBR
      would be configured as an FIR.
   RIB/FIB-Suppressing Router (FSR):  An S-VA router that installs a
      route to 0/0, and may suppress other routes.  Typically an edge
      router would be configured as an FSR.

   Install and Suppress:  The terms "install" and "suppress" are used to
      describe whether a protocol local RIB entry has been loaded or not
      loaded into the global RIB and FIB.  In other words, the phrase
      "install a route" means "install a route into the global RIB and
      FIB", and the phrase "suppress a route" means "do not install a
      route from BGP into the global RIB and FIB".
   Legacy Router:  A router that does not run S-VA, and has no knowledge
      of S-VA.
   Global Routing Information Base (RIB):  The term global RIB is used
      to indicate the router's main routing information base.  That RIB
      is normally used to populate FIB tables of the router.  It needs
      to be highlighted that unless FIB compression is used global RIB
      and FIB tables are in sync.
   Local/Protocol Routing Information Base (loc-RIB):  The term local
      RIB is used to indicate the protocol's table where product of SPF
      or BGP best path selection is kept before being installed in
      global RIB.  In some protocol's implementations for example BGP
      loc-RIB can be further divided into Adj-RIBs-In, the Loc-RIB, and
      the Adj-RIBs-Out.

2.  Operation of S-VA

   There are three types of routers in S-VA, FIB-Installing routers
   (FIR), FIB-Suppressing routers (FSR), and optionally legacy routers.
   While any router can be an FIR or an FSR (there are no topology
   constraints), the most simple form of deployment is for AS border or
   POP border routers to be configured as FIRs, and for customer facing
   edge routers respectively in the AS or in the POP to be configured as
   FSRs.

   FIRs must originate a default BGP route to NLRI 0/0 [RFC4271].  The
   ORIGIN is set to INCOMPLETE (value 2) and the BGP NEXT_HOP is set to
   match the other BGP routes which are also advertised by said FIR.
   The ATOMIC_AGGREGATE and AGGREGATOR attributes are not included.  The
   FIR MUST attach a NO_EXPORT Community Attribute [RFC1997] to the
   default route.

   FIRs should not FIB-suppress any routes.  They may, however, still
   use some form of local FIB compression algorithm if deemed necessary.

   FSRs must detect the VA prefix 0/0 and install it both in loc-RIB,
   RIB and FIB.  Following that FSR may suppres any more specific routes
   which carry the same next hop as the VA prefix.  To guarantee
   semantical correctness FSR by default should also be able to detect
   installation of not matching next hop route and reinstall all the
   more specifics which were previously eligible for suppression to
   maintain semantical forwarding correctness.

   Generally, any more specific route which carries the same next hop as
   the VA-prefix 0/0 is eligible for suppression.  However, provided
   that there was at least one less specific prefix (e.g., 1.0.0.0/8)
   and the next-hop of such prefix was different from that of the VA
   0/0, those more specific prefixes (e.g., 1.1.1.0/24) which are
   otherwise subject to suppression would not be eligible for
   suppression anymore.

   Similarly when IBGP multipath is enabled and when multiple VA
   prefixes are detected which are multipath candidates under given
   network condition only those more specific prefixes are subject to
   suppresion which have the identical set of next hops as multipath set
   of VA prefixes.

   Let's illustrate the expected behaviour on the figure below.  This
   figure shows an autonomous system with a FIR FIR1 and an FSR FSR1.
   FSR1 is an ASBR and is connected to two remote ASBRs, EP1 and EP2.

        +------------------------------------------+
        |      Autonomous System                   |   +----+
        |                                          |   |EP1 |
        |                                      /---+---|    |
        |   To   ----\ +----+          +----+ /    |   +----+
        | Other       \|FIR1|----------|FSR1|/     |
        |Routers      /|    |          |    |\     |
        |        ----/ +----+          +----+ \    |   +----+
        |                                      \---+---|EP2 |
        |                                          |   |    |
        |                                          |   +----+
        +------------------------------------------+

   Suppose that FSR1 has been enabled to perform S-VA.  Originally it
   receives all routes from FIR1 (doing next hop self) as well as
   directly connected EBGP peers EP1 and EP2.  FIR1 now will advertise a
   VA prefix 0/0 with next hop set to himself.  That will trigger
   detection of such prefix on FSR1 and suppression all routes which
   have the same next hop as VA prefix and which otherwise would be
   installed in RIB and FIB.  However it needs to be observed that FSR1
   will not suppres any EBGP routes received from his peers EP1 and EP2
   due to next hop being different from the one assinged to VA-prefix.

3.  Deployment considerations

   The simplest deployment model of S-VA is it's use within the POP.  In
   such model the POP to core boundary routers (usually RRs in the data
   path) would act as FIRs and would inject VA-prefix 0/0 to all of it's
   clients within the POP.  In such model of operation an observation
   can be made that such ABRs do have full routing knowledge and client
   to ABR distance is negligable as compared with client to intra-domain
   exit distance.

   Therefor under the above intra POP S-VA deployment model clients can
   be configured that even in the event of lack of ABR to ABR
   advertisement symmetry there is still no need to monitor if more
   specific unsuppressed route would cover suppressed one.  Thus in this
   particular deployment model there is no need to detect and reinstall
   the previously suppressed ones.

   Another deploymet consideration should be given to networks which may
   utilize route reflection.  In the event of enabling IBGP multipath a
   special care must be taken that both outbound prefixes as well as VA-
   prefixes would pass via said route reflectors to their clients.

   In order to addess the above aspects the following solutions could be
   considered:

   - Use of intra-POP S-VA
   - Full mesh  Small or medium side networks where S-VA can be deployed
      are normally fully meshed and do not use route reflection.  It
      also needs to pointed out that some large networks are also fully
      meshed today.
   - Use of add-paths  Use of add-paths new BGP encoding will allow to
      distribute more then one overall best path from RR to each client.
   - Alternate advertisement of VA-prefix   S-VA prefix does not need to
      be advertised in BGP.  The BGP suppression will happen as long as
      we configure the S-VA with next hop(s) and implementation verifies
      that such VA-prefix is installed in the RIB and FIB.

4.  IANA Considerations

   There are no IANA considerations.

5.  Security Considerations

   The authors are not aware of any new security considerations due to
   S-VA.

6.  Acknowledgements

   The concept for Virtual Aggregation comes from Paul Francis.  In this
   document authors only simplified some aspects of it's behaviour to
   allow simpler adoption by some operators.

   Authors would like to thank Clarence Filsfils for his valuable input.

7.  References

7.1.  Normative References

   [RFC1997]  Chandrasekeran, R., Traina, P., and T. Li, "BGP
              Communities Attribute", RFC 1997, August 1996.

   [RFC2119]  Bradner, S., "Key words for use in RFCs to Indicate
              Requirement Levels", BCP 14, RFC 2119, March 1997.

   [RFC4271]  Rekhter, Y., Li, T., and S. Hares, "A Border Gateway
              Protocol 4 (BGP-4)", RFC 4271, January 2006.

7.2.  Informative References

   [I-D.ietf-grow-va]
              Francis, P., Xu, X., Ballani, H., Jen, D., Raszuk, R., and
              L. Zhang, "FIB Suppression with Virtual Aggregation",
              draft-ietf-grow-va-05
              draft-ietf-grow-va-06 (work in progress), June December 2011.

Authors' Addresses

   Robert Raszuk
   NTT MCL
   101 S Ellsworth Avenue Suite 350
   San Mateo, CA  94401
   US

   Email: robert@raszuk.net

   Alton Lo
   Cisco Systems, Inc.
   170 West Tasman Drive
   San Jose,
   Arista Networks
   5470 Great America Parkway
   Santa Clara, CA  95134  95054
   USA

   Phone:
   Email: altonlo@cisco.com altonlo@aristanetworks.com
   Lixia Zhang
   UCLA
   3713 Boelter Hall
   Los Angeles, CA  90095
   US

   Phone:
   Email: lixia@cs.ucla.edu

   Xiaohu Xu
   Huawei Technologies
   No.3 Xinxi Rd., Shang-Di Information Industry Base, Hai-Dian District
   Beijing, Beijing  100085
   P.R.China

   Phone: +86 10 82836073
   Email: xuxh@huawei.com