rfc2791.txt

来自「RFC 的详细文档！」· 文本 代码 · 共 1,460 行 · 第 1/5 页

Network Working Group                                              J. Yu
Request for Comments: 2791                         CoSine Communications
Category: Informational                                        July 2000


                   Scalable Routing Design Principles

Status of this Memo

   This memo provides information for the Internet community.  It does
   not specify an Internet standard of any kind.  Distribution of this
   memo is unlimited.

Copyright Notice

   Copyright (C) The Internet Society (2000).  All Rights Reserved.

Abstract

   Routing is essential to a network. Routing scalability is essential
   to a large network. When routing does not scale, there is a direct
   impact on the stability and performance of a network. Therefore,
   routing scalability is an important issue, especially for a large
   network. This document identifies major factors affecting routing
   scalability as well as basic principles of designing scalable routing
   for large networks.

























Yu                           Informational                      [Page 1]

RFC 2791           Scalable Routing Design Principles          July 2000


Table of Contents

   1           Introduction  ..................................      2
   2           Common Routing Design Goals  ...................      3
   3           Characteristics of Today's Large Networks  .....      3
   4           Routing Scaling Issues  ..........................    3
   4.1         Router Resource Consumption  .....................    4
   4.2         Routing Complexity  ..............................    5
   5           Routing Protocol Scalability .....................    6
   5.1         IS-IS and OSPF  ..................................    6
   5.2         BGP  .............................................    8
   6           Scalable Routing Design Principles  ..............    9
   6.1         Building Hierarchy  ..............................   10
   6.2         Compartmentalization  ............................   13
   6.3         Making Proper Trade-offs  ........................   13
   6.4         Reduce Burdens of Routing Information Process  ...   14
   6.4.1       Routing Intelligence Placement  ..................   14
   6.4.2       Reduce Routes and Routing Information  ...........   15
   6.4.2.1     CIDR and Route Aggregation  ......................   15
   6.4.2.2     Utilize Default Routing where it's Possible  .....   15
   6.4.2.3     Reduce Alternative Paths  ........................   16
   6.4.3       Use Static Route at Edge  .........................  16
   6.4.4       Minimize the Impact of Route Flapping  ............  16
   6.5         Scalable Routing Policy and Scalable Implementation  17
   6.6         Out-of-band Process  ..............................  19
   7           Conclusion and Discussion  ........................  19
   8           Security Considerations  ..........................  20
   9           Acknowledgement  ..................................  21
   10          References  .......................................  21
   Author's Address ..............................................  22
   Appendix A  Out-of-Band Routing Processes  ....................  23
   Full Copyright Statement  .....................................  26

1. Introduction

   Routing is essential to a network. Without routing, packets cannot be
   delivered to desired destinations and the network would be non-
   functional. The challenge of designing the routing for a large
   network, such as a large ISP backbone network, is not only to make it
   work, but also to make it scale. Without a scalable routing system, a
   network may suffer from severe performance penalties, as
   unfortunately proven by disastrous events in large networks. This
   document attempts to analyze routing scalability issues and define a
   set of principles for designing scalable routing system for large
   networks.

   The organization of this document is as follows: Section 2 describes
   routing functions and design goals. Sections 3 and 4 discuss the



Yu                           Informational                      [Page 2]

RFC 2791           Scalable Routing Design Principles          July 2000


   characteristics of today's large networks and the associated routing
   scaling issues. Section 5 explores routing protocol scalability, and
   Section 6 presents scalable routing design principles. Section 7
   provides a conclusion to the document.

2. Common Routing Design Goals

   The basic goals a routing system should achieve are as follows:

      o Stability
      o Redundancy and robustness
      o Reasonable convergency time
      o Routing information integrity
      o Sensible and manageable routing policy

   The challenge of designing routing in a large network is not only to
   achieve these basic goals but also to make the routing system scale.

3. Characteristics of Today's Large Networks

   Today's large networks typically possess the following features:

      o They are composed of a large number of nodes (routers and/or
        switches), typically in the hundreds. Some provider networks
        include customer CPE routers within their administrative domain,
        which increases the number of nodes to thousands.

      o They have rich connectivity to meet redundancy and robustness
        requirements, and they consequently have complex topologies.

      o They are default-free; that is, they carry all the routes known
        to the entire Internet. Currently, the total number is
        approximately 70,000.

      o The customer aggregation routers inside the large networks
        connect sometimes hundreds of customer routers.

   These characteristics impose a direct challenge to the routing
   scalability of the network.

4. Routing Scaling Issues

   Today, the main issues surrounding routing scaling are: i) excessive
   router resource consumption, which can potentially increase routing
   convergency difficulties thus destabilize a network; and ii) routing
   complexity, resulting in poor management of network, producing low
   service quality.




Yu                           Informational                      [Page 3]

RFC 2791           Scalable Routing Design Principles          July 2000


4.1. Router Resource Consumption

   The routing process puts bursty loads on routers, especially under
   unstable network conditions. In the extreme case, the routing process
   takes all available resources from the routers, which results in slow
   routing convergence or no convergence. A network is paralyzed when it
   cannot converge internal routing information.

   It's worthy noting that routers with internal architectures that
   tightly couple forwarding and routing processes tend to handle the
   excessive routing load poorly. The emerging new generation of routers
   with the architecture of separating resource used for forwarding and
   routing could provide better routing scalability.

   Today, a large network typically employs IS-IS [1,2] or OSPF [3] as
   an Interior Routing Protocol(IGP) and BGP [4] as an Exterior Routing
   Protocol(EGP), respectively. The IGP calculates paths across the
   interior of the network. BGP facilitates routing exchange between
   routing domains, or Autonomous Systems (AS). BGP also processes and
   propagates external routing information within the network. The
   presence of a large number of routers and adjacencies in a network,
   coupled with frequent topology changes due to link instability, will
   contribute to excessive resource consumption by the interior routing.
   In the case of exterior routing, a large quantity of routers in a BGP
   system plus frequent routing updates (route flapping) would put a
   heavy burden on the routers. Section 5 describes scaling issues with
   IS-IS, OSPF and BGP in detail.

   In addition, having many destinations in a routing system, combined
   with multiple paths associated with these routes, impose the
   following scaling issues on BGP:

      o A large number of routes combined with multiple paths for each
        increases the cost of routing processing for route selection,
        routing policy application and filtering.

      o Too many routes combined with multiple paths requires large
        amounts of memory on routers for storage. The demand is even
        higher at InterExchange Points such as NAPs.

      o The larger the number of routes, the greater the chance route
        flapping will occur and the more BGP routing updates will happen
        as a result. Based on statistics collected by [5], thousands of
        BGP updates in a measured 15 minute interval can occur on a
        typical default-free router at a NAP.






Yu                           Informational                      [Page 4]

RFC 2791           Scalable Routing Design Principles          July 2000


        Route flapping refers to frequent routing updates occurring due
        to network instability, for example, when the state of a
        physical link in the network is fluctuating, or when a BGP
        session is torn down and re-established numerous time within a
        short period of time.

        To facilitate fast convergence, topology change information must
        be propagated in a timely fashion. When a route becomes
        unavailable and is withdrawn, the information is typically sent
        immediately. If the affected routes have been announced to the
        global Internet, the update information is likely to be
        propagated to the entire Internet.

        Route flapping has a profound impact on routers running BGP. The
        routers have to process routing information frequently and this
        consumes a tremendous amounts of the available resources. When a
        local route or link is oscillating, interior routing is affected
        as well by excessive topology information flooding and
        subsequent shortest path calculations. However, OSPF (or IS-IS)
        imposes rate limits on such activity to reduce the burden on the
        routers. For example, OSPF specifies that an individual SLA can
        be updated at most once every 5 seconds. This essentially
        dampens the flapping.

   Moreover, large numbers of E-BGP sessions processed by a single
   router create another potential scaling issue. Large networks usually
   have huge customer subscriptions and connections. To scale the
   hardware and the number of nodes in the network, providers tend to
   dedicate a group of customer aggregation routers, each connecting as
   many customer CPE routers as possible. As a result, it's not uncommon
   for a customer aggregation router to handle hundreds of E-BGP
   sessions, which imposes potential problems, such as BGP session
   processing and maintenance, route processing, filtering and route
   storage.

4.2. Routing Complexity

   Routing complexity can lead to network management difficulties, which
   will have an impact on trouble shooting and quick problem resolution.
   It can result in a less than desirable service quality across the
   network. Complicated routing policies and special cases or exceptions
   in a routing design can contribute to routing complexity in a large
   system.

   Routing Policy refers to the administrative criteria for handling
   routing information, commonly in the form of routing path selection
   and route filtering. The way routing information is handled has a
   direct impact on traffic flow within a network and across domains. As



Yu                           Informational                      [Page 5]

RFC 2791           Scalable Routing Design Principles          July 2000


   a result, it affects business agreements among different networks.
   Therefore, the determination of routing policy is largely dominated
   by non-technical concerns, such as business considerations. Routing
   policy can be very complex, which would make management and
   configuration an unscalable task.