Network Working Group Request for Comments: 1998 Category: Informational cisco Systems August 1996

Similar documents
Network Working Group Request for Comments: 2519 Category: Informational Juniper February A Framework for Inter-Domain Route Aggregation

Request for Comments: T. Li August 1996

Network Working Group. Category: Standards Track Cisco Systems. D. Katz Juniper Networks Y. Rekhter. Cisco Systems. February 1998

Network Working Group. J. Scudder Cisco Systems, Inc. February 2001

Internet Engineering Task Force (IETF) Category: Standards Track. Cisco Systems, Inc. J. Scudder Juniper Networks September 2016

Network Working Group. Category: Experimental June 1996

Network Working Group Request for Comments: August Address-Prefix-Based Outbound Route Filter for BGP-4

Balancing incoming traffic over multiple links

Connecting to a Service Provider Using External BGP

L11 : Inter-domain Routing with BGP Lecture14 Michaelmas, 2016

Request for Comments: 3345 Category: Informational AOL Time Warner, Inc. D. Walton A. Retana Cisco Systems, Inc. August 2002

BGP. Autonomous system (AS) BGP version 4

BGP. Autonomous system (AS) BGP version 4. Definition (AS Autonomous System)

BGP. Autonomous system (AS) BGP version 4

BGP. Autonomous system (AS) BGP version 4. Definition (AS Autonomous System)

BGP. Autonomous system (AS) BGP version 4

Border Gateway Protocol (an introduction) Karst Koymans. Monday, March 10, 2014

BGP Attributes (C) Herbert Haas 2005/03/11 1

ISP Border Definition. Alexander Azimov

Network Working Group. Category: Standards Track Juniper Networks August 2008

BGP Attributes and Path Selection

TELE 301 Network Management

Lab 2 BGP route filtering and advanced features

Module 8 Multihoming Strategies Lab

Category: Standards Track Inria March Use of BGP-4 Multiprotocol Extensions for IPv6 Inter-Domain Routing

Internet Routing Protocols Lecture 03 Inter-domain Routing

BGP Policy Accounting Output Interface Accounting

Multihoming Techniques. bdnog8 May 4 8, 2018 Jashore, Bangladesh.

Network Working Group. Category: Standards Track Cisco Systems May 2007

Brent Sweeny Global Research NOC at Indiana University (USA) Terena Network Conference 19 April 2014 (Dublin)

Internet Engineering Task Force (IETF) Request for Comments: 6368 Category: Standards Track

Internet Engineering Task Force (IETF) Category: Standards Track December 2012 ISSN:

Introduction to IP Routing. Geoff Huston

From the given configuration taken from RTA and graphic, which network will be filtered from being propagated to RTC from RTA?

IPv4/IPv6 BGP Routing Workshop. Organized by:

Network Working Group. Redback H. Smit. Procket Networks. October Domain-wide Prefix Distribution with Two-Level IS-IS

Network Working Group. Obsoletes: 1656 March 1995 Category: Informational

BGP made easy. John van Oppen Spectrum Networks / AS11404

Module 18 Transit. Objective: To investigate methods for providing transit services. Prerequisites: Modules 12 and 13, and the Transit Presentation

BGP Policy Accounting

Connecting to a Service Provider Using External BGP

BGP and the Internet

Cisco CISCO Configuring BGP on Cisco Routers Exam. Practice Test. Version

Module 12 Multihoming to the Same ISP

LARGE SCALE IP ROUTING LECTURE BY SEBASTIAN GRAF

Internet Engineering Task Force (IETF) Request for Comments: 6769 Category: Informational. A. Lo Arista L. Zhang UCLA X. Xu Huawei October 2012

BGP. Autonomous system (AS) BGP version 4. Definition (AS Autonomous System)

BGP Graceful Shutdown

February Copyright (c) 2009 IETF Trust and the persons identified as the document authors. All rights reserved.

BGP Cost Community. Prerequisites for the BGP Cost Community Feature

BGP. Border Gateway Protocol (an introduction) Karst Koymans. Informatics Institute University of Amsterdam. (version 17.3, 2017/12/04 13:20:08)

BGP Policy Control. ISP Workshops

BGP. Border Gateway Protocol A short introduction. Karst Koymans. Informatics Institute University of Amsterdam. (version 18.3, 2018/12/03 13:53:22)

Internet Engineering Task Force (IETF) Request for Comments: Google K. Patel Cisco Systems August 2015

BGP Configuration. BGP Overview. Introduction to BGP. Formats of BGP Messages. Header

Module 3 BGP route filtering and advanced features

TBGP: A more scalable and functional BGP. Paul Francis Jan. 2004

BGP. Attributes 2005/03/11. (C) Herbert Haas

Lab 3 Multihoming to the Same ISP

BGP Configuration for a Transit ISP

Routing and RFC AKA using BGP communities to influence routing

BGP Policy Control. ISP Workshops. Last updated 17 May 2014

BGP Multihoming ISP/IXP Workshops

BGP Attributes and Policy Control

Service Provider Multihoming

Lecture 18: Border Gateway Protocol

BGP and the Internet

BGP. Internal and External BGP 2005/03/11. (C) Herbert Haas

BGP Attributes and Policy Control

CS BGP v4. Fall 2014

BGP Attributes and Policy Control

Lecture 17: Border Gateway Protocol

Shim6: Network Operator Concerns. Jason Schiller Senior Internet Network Engineer IP Core Infrastructure Engineering UUNET / MCI

BGP and the Internet. Why Multihome? Why Multihome? Why Multihome? Why Multihome? Why Multihome? Redundancy. Reliability

Module 13 Multihoming to Different ISPs

Tag Switching. Background. Tag-Switching Architecture. Forwarding Component CHAPTER

Introduction to BGP ISP/IXP Workshops

BGP. Autonomous system (AS) BGP version 4. Definition (AS Autonomous System)

Border Gateway Protocol (an introduction) Karst Koymans. Tuesday, March 8, 2016

This appendix contains supplementary Border Gateway Protocol (BGP) information and covers the following topics:

Service Provider Multihoming

Opaque Information Distribution

BGP-v4 Theory and Practice

LACNIC XIII. Using BGP for Traffic Engineering in an ISP

EE 122: Inter-domain routing Border Gateway Protocol (BGP)

BGP route filtering and advanced features

MPLS VPN Carrier Supporting Carrier IPv4 BGP Label Distribution

Alcatel-Lucent 4A Alcatel-Lucent Border Gateway Protocol. Download Full Version :

CS 640: Introduction to Computer Networks. Intra-domain routing. Inter-domain Routing: Hierarchy. Aditya Akella

Internet Routing Protocols Lecture 01 & 02

Module 2 More ibgp, and Basic ebgp Configuration

University of Belgrade - School of Electrical Engineering Department of Telecommunications

Outline. Organization of the global Internet. BGP basics Routing policies The Border Gateway Protocol How to prefer some routes over others

Internet Protocols Fall Lectures Inter-domain routing, mobility support, multicast routing Andreas Terzis

Lecture 16: Border Gateway Protocol

Outline Computer Networking. Inter and Intra-Domain Routing. Internet s Area Hierarchy Routing hierarchy. Internet structure

Ravi Chandra cisco Systems Cisco Systems Confidential

by Jun ichi Endo * 1. INTRODUCTION

Internet Interconnection Structure

BGP and the Internet. Enterprise Multihoming. Enterprise Multihoming. Medium/Large ISP Multihoming. Enterprise Multihoming. Enterprise Multihoming

Transcription:

Network Working Group Request for Comments: 1998 Category: Informational E. Chen MCI T. Bates cisco Systems August 1996 An Application of the BGP Community Attribute in Multi-home Routing Status of This Memo This memo provides information for the Internet community. This memo does not specify an Internet standard of any kind. Distribution of this memo is unlimited. Abstract This document presents an application of the BGP community attribute [2] in simplifying the implementation and configuration of routing policies in the multi-provider Internet. It shows how the community based configuration can be used to replace the AS-based customization of the BGP "LOCAL_PREF" attribute, a common method used today. Not only does the technique presented simplifies configuration and management at the provider level, it also represents a paradigm shift in that it gives the potential for the customer to control its own routing policy with respect to its service provider, as well as providing the ability for policy configuration to be done at a prefix based granularity rather than the more common AS based granularity. 1. Introduction In the multi-provider Internet, it is common for a service subscriber (i.e., customer) to have more than one service provider, or to have arrangements for redundant connectivity to the global connected Internet. As discussed in [3], routing strategies in these cases usually require coordination between the service subscriber and its providers, which typically leads to customization of router configurations (e.g., BGP "LOCAL_PREF") not only by the subscriber, but also by its providers. Due to the large number of customers a provider serves, customization of router configurations at the provider level may present management and scalability problems. This document presents an application of the BGP community attribute in simplifying the implementation of routing strategies in the multi-provider Internet. More specifically, the technique presented uses a community-based, rather than the common AS-based, Chen & Bates Informational [Page 1]

configuration of the BGP "LOCAL_PREF". It essentially removes the need for customized configuration of the BGP "LOCAL_PREF" attribute at the provider level while maintaining the same level of routing functionality and flexibility. It also represents a paradigm shift in that it gives the potential for the customer to control its own routing policy with respect to its service provider, as well as providing the ability for policy configuration to be done at a prefix based granularity rather than the more common AS based granularity in use today. 2. AS-based Configuration and its Drawbacks As discussed in [3], in today s multi-provider Internet, customized configuration of the BGP "LOCAL_PREF" attribute is often required to implement common routing strategies such as load-sharing or backup. There are two main reasons: o Lack of available implementations and deployment of routing software that supports the "Destination Preference Attribute" (DPA) as specified in [4]. DPA allows one to specify a globally transitive preference so that return traffic favors certain path. As discussed in [3], the attribute will be very useful in influencing route selection for routes with identical "LOCAL_PREF" and equal AS-path length. o In the multi-provider Internet, it is common for a provider to assign higher BGP "LOCAL_PREF" values for routes from its customers than from other service providers. This practice provides some degree of protection for its customer routes, and it facilitates implementation of certain routing strategies. It, however, also complicates other routing implementations such as backup arrangement, thus, requiring customized "LOCAL_PREF" configuration. Figure 1 shows a typical case of a backup arrangement in the multiprovider Internet. In Figure 1, AS1 and AS2 are both providers, and AS3 and AS4 are customers of AS1 and AS2, respectively. AS3 has entered a bilateral agreement with AS4 to provide backup to each other. That is, AS3 would use its direct link to AS4 to reach only AS4 in the normal circumstance, and for transit in the case of a failure between AS3 and AS1. To realize this routing agreement, AS3 requests that its provider AS1 adjust its BGP "LOCAL_PREF" configuration so that AS1 reaches AS4 via AS2. Chen & Bates Informational [Page 2]

+------+ +------+ AS1 ------ AS2 +------+ +------+ +------+ +------+ AS3 ------ AS4 +------+ +------+ Figure 1: Typical Backup Scenario Primarily due to scalability and management concerns, most providers only perform "LOCAL_PREF" customization based on ASs, not on IP prefixes. If IP prefix-based "LOCAL_PREF" configuration is needed, a technique known as as the BGP AS-path manipulation can be used. However, it is currently only available in certain vendor s products. There are several drawbacks with the the practice of AS-based BGP "LOCAL_PREF" configuration at the provider level: o The implementation tends to less efficient due to the process of coordination and configuration. More importantly, the process needs to be repeated each time a change (e.g., adding a new AS) occurs. o The AS-based customization complicates router configuration and increases complexity of network operation. It has become a serious scalability issue for providers. o It can not implement prefix-based configuration without the AS-path manipulation (i.e., using fake AS). o Keeping configuration up-to-date is some times problematic. 3. How the BGP Community Attribute Can Help 3.1 Overview of the Community Attribute The BGP community path attribute is an optional transitive attribute of variable length [1,2]. The attribute consists of a set of four octet values, each of which specify a community. The community attribute values are encoded using an AS number in the first two octets, with the remaining two octets defined by the AS. As defined in [2], a community is a group of destinations (i.e. prefixes) that share some common attribute. Each destination can belong to multiple communities. All prefixes with the community attribute belong to the communities listed in the attribute. Chen & Bates Informational [Page 3]

The BGP community allows one to group a set of prefixes and perform routing decisions based on the identity of the group. The well-known communities NO_EXPORT (0xFFFFFF01) and NO_ADVERTISE (0xFFFFFF02) are intuitive, and can be used for optimizing routing and for improving route aggregation. 3.2 Community-based Configuration With the BGP community attribute [2], a provider can now use community-based, rather than AS-based, configuration of BGP "LOCAL_PREF". The provider first needs to coordinate with its customers a set of communities to be mapped to certain BGP "LOCAL_PREF" values. The provider can then apply a uniform BGP configuration to all its customers that would capture routes with the community values, and set up the appropriate BGP "LOCAL_PREF" values accordingly. A customer that requires customization in its provider BGP "LOCAL_PREF" configuration can simply send the appropriate community values in its routing announcements. The major advantages of using this technique include: o The customer has full control in the process, which makes a lot of sense as the customer is in a position to have better understanding about its own topology and routing policy requirement. o The effect of route-based customization in BGP "LOCAL_PREF" configuration by providers can now be achieved, thus, removing the need of AS-Path manipulation in certain cases. o It addresses the scalability issue facing providers as it distributes the configuration work to the customer that requires customization. Chen & Bates Informational [Page 4]

4. A Real-World Implementation Example MCI currently makes heavy use of the BGP "LOCAL_PREF" attribute value as part of its routing policy configuration process. Different BGP "LOCAL_PREF" values are assigned for routes from different sources. Table 1 details these values: +-------------------------+------------+ Category LOCAL_PREF +-------------------------+------------+ Customer Routes 100 Customer backup Routes 90 Other ISP Routes 80 Customer-Provided backup 70 +-------------------------+------------+ Table 1: Defined LOCAL_PREF Values Note: o The value 100 is the default value used within our network configuration. o In most cases, the MED attribute set by a customer is sufficient for customer backup routes (e.g., T1 backs up T3). However, in certain cases configuration of "LOCAL_PREF" will still be necessary until the BGP DPA attribute is available. To make use of the BGP community attribute, several community values (MCI s AS number: 3561 = 0x0DE9) have been defined that can be used by customers to tag routes so that the appropriate "LOCAL_PREF" values are configured. Table 2 lists the appropriate community attribute values (and the mappings of community to LOCAL_PREF): +---------------------+------------+ community LOCAL_PREF +---------------------+------------+ 3561:70 (0x0DE90046) 70 3561:80 (0x0DE90050) 80 3561:90 (0x0DE9005A) 90 +---------------------+------------+ Table 2: Community to LOCAL_PREF Mapping Chen & Bates Informational [Page 5]

A customer requiring MCI to configure BGP "LOCAL_PREF" values other than the default can tag their routes with the defined communities. The community values can be configured either based on an AS path list or an IP address access list. A cisco systems software specific configuration example is given in Appendix A to show how this can be achieved. A uniform BGP configuration (see Appendix B, again cisco systems software specific) is applied by MCI to peers with customers that configure the appropriate "LOCAL_PREF" values based on the communities received. This technique has been tested and is in use with several customers, and the response has been very positive. We are in the process of migrating all other customized BGP "LOCAL_PREF" configurations to this uniform community based configuration approach. 5. References [1] Rekhter, Y., and Li, T., "A Border Gateway Protocol 4 (BGP-4)", RFC 1771, March 1995. [2] Chandra, R., Traina, P., and T. Li, "BGP Communities Attribute", RFC 1997, August 1996. [3] Chen, E., and T. Bates, "Current Practice of Implementing Symmetric Routing and Load Sharing in the Multi-Provider Internet", Work in Progress. [4] Chen, E., and T. Bates, "Destination Preference Attribute for BGP", Work in Progress. [5] Chen, E., and T. Bates, "Application of the BGP Destination Preference Attribute in Implementing Symmetric Routing", Work in Progress. [6] cisco systems, cisco IOS Software Version 10.3 Router Products Configuration Guide (Addendum), May 1995. 6. Security Considerations Security issues are not discussed in this memo. 7. Acknowledgments The authors would specifically like to thank Ravi Chandra, Tony Li and Paul Traina of cisco systems for devising and implementing the community attribute. Chen & Bates Informational [Page 6]

8. Authors Addresses Enke Chen MCI 2100 Reston Parkway Reston, VA 22091 Phone: +1 703 715 7087 EMail: enke@mci.net Tony Bates cisco Systems 170 West Tasman Drive San Jose, CA 95134 Phone: +1 408 527 2470 EMail: tbates@cisco.com Chen & Bates Informational [Page 7]

Appendix These appendices list cisco systems software specific configuration examples for configuring communities, and for uniform route-map definition that sets up the appropriate "LOCAL_PREF" values based on the corresponding community values. These examples are given purely to show a working example of how the desired effect discussed in this document can be achieved. Please refer to [6] for more specific information on cisco configuration and syntax. Appendix A. Community Configuration The community values can be configured either based upon an AS path list or based an IP address access list. Here is an example that includes both cases: router bgp xxxx neighbor x.x.x.x remote-as 3561 neighbor x.x.x.x filter-list 20 out neighbor x.x.x.x route-map config-community out neighbor x.x.x.x send-community # match all ip as-path access-list 1 permit.* # list of customer ASs ip as-path access-list 20 permit ^$ ip as-path access-list 20 permit ^64700_ ip as-path access-list 20 deny.* # AS path based matching, backup for another ISPs customer ip as-path access-list 40 permit _64710_ ip as-path access-list 40 permit _64711_ ip as-path access-list 40 deny.* # route-map route-map config-community permit 10 match as-path 40 set community 0x0DE90046 route-map config-community permit 20 match as-path 1 Chen & Bates Informational [Page 8]

Note: The community can also be configured based on IP prefixes instead of AS numbers. For example, access-list 101 permit ip 192.160.154.0 0.0.0.0 255.255.255.0 0.0.0.0 route-map config-community permit 10 match ip address 101 set community 0x0DE90046 route-map config-community permit 20 match as-path 1 Appendix B. Uniform Route-map Configuration Here is the uniform route-map that can be used for all BGP customers: # routes primary via another ISP ip community-list 70 permit 0x0DE90046 ip community-list 70 deny # routes also homed to another ISP, but with DPA or # AS-path length as the tie-breaker ip community-list 80 permit 0x0DE90050 ip community-list 80 deny # customer backup routes ip community-list 90 permit 0x0DE9005A ip community-list 90 deny # the route-map applied to BGP customers route-map set-customer-local-pref permit 10 match community 70 set local-preference 70 route-map set-customer-local-pref permit 20 match community 80 set local-preference 80 route-map set-customer-local-pref permit 30 match community 90 set local-preference 90 route-map set-customer-local-pref permit 40 match as-path 1 set local-preference 100 Chen & Bates Informational [Page 9]