1

An Introduction to Interdomain Routing

and BGP

Timothy G. Griffin

griffin@research.att.com

http://www.research.att.com/~griffin/interdomain.html

SIGCOMM 2001 Tutorial Session

August 28, 2001

Acknowledgements

Thanks to Jay Borkenhagen, Randy Bush, Anja Feldmann, Matt Grossglauser, Madan Musuvathi, Jennifer Rexford, Shubho Sen, and Jia Wang for many helpful comments

My opinions should not be taken torepresent AT&T policy

Errors are my own

Common View of the Telco Network

Brick

Common View of the IP Network (Layer 3)

What This Tutorial Is About

routing

GoalUnderstand how layer 3 connectivity is maintained in the global Internet

This tutorial will not say much about the applications that exploit this connectivity.

It will be restricted to IPv4 unicast routing.

• Part I : The basics of interdomain routing and BGP

• Part II : BGP in practice: Issues of Scale

Outline Part I

• Forwarding vs. Routing• IP addressing• Autonomous Systems (basic units of

interdomain routing)• The Border Gateway Protocol (BGP)

– BGP fundamentals – BGP route attributes– Implementing policy with BGP – A wee bit of theory

Outline Part II

• Scaling internal BGP• BGP table growth

– Address aggregation vs. Multihoming

• Growth in number of autonomous systems

• Dynamics of BGP– Route flapping – BGP convergence– Rates of BGP updates

Best Effort Connectivity

This is the fundamental service provided by Internet Service Providers (ISPs)

All other IP services depend on connectivity: DNS, email, VPNs, Web Hosting, …

IP traffic

135.207.49.8 192.0.2.153

10

Routing vs. Forwarding

R

R

RA

B

C

D

R1

R2

R3

R4 R5

ENet Nxt Hop

R4R3R3R4DirectR4

Net Nxt Hop

A B C D Edefault

R2R2DirectR5R5R2

Net Nxt Hop

A B C D Edefault

R1DirectR3R1R3R1

Default toupstreamrouter

A B C D Edefault

Forwarding: determine next hop

Routing: establish end-to-end paths

Forwarding always works

Routing can be badly broken

11

How Are Forwarding Tables Populated to implement

Routing?Statically Dynamically

Routers exchange network reachability information using ROUTING PROTOCOLS. Routers use this to compute best routes

Administrator manually configuresforwarding table entries

In practice : a mix of these.Static routing mostly at the “edge”

+ More control+ Not restricted to destination-based forwarding - Doesn’t scale- Slow to adapt to network failures

+ Can rapidly adapt to changes in network topology+ Can be made to scale well- Complex distributed algorithms- Consume CPU, Bandwidth, Memory- Debugging can be difficult- Current protocols are destination-based

Routers Talking to Routers

Routing info

Routing info

• Routing computation is distributed among routers within a routing domain

• Computation of best next hop based on routing information is the most CPU/memory intensive task on a router

• Routing messages are usually not routed, but exchanged via layer 2 between physically adjacent routers (internal BGP and multi-hop external BGP are exceptions)

Before We Go Any Further …

IP ROUTING PROTOCOLS DO NOT DYNAMICALLY ROUTE AROUND NETWORK CONGESTION

• IP traffic can be very bursty • Dynamic adjustments in routing typically

operate more slowly than fluctuations in traffic load

• Dynamically adapting routing to account for traffic load can lead to wild, unstable oscillations of routing system

Autonomous Routing Domains

A collection of physical networks glued togetherusing IP, that have a unified administrativerouting policy.

• Campus networks• Corporate networks• ISP Internal networks• …

Autonomous Systems (ASes)

An autonomous system is an autonomous routing domainthat has been assigned an Autonomous System Number (ASN).

RFC 1930: Guidelines for creation, selection, and registration of an Autonomous System

… the administration of an AS appears to other ASes to have a single coherent interior routing plan and presents a consistent picture of what networks are reachable through it.

AS Numbers (ASNs)ASNs are 16 bit values.

64512 through 65535 are “private”

• Genuity: 1 • MIT: 3• Harvard: 11• UC San Diego: 7377• AT&T: 7018, 6341, 5074, … • UUNET: 701, 702, 284, 12199, …• Sprint: 1239, 1240, 6211, 6242, …• …

ASNs represent units of routing policy

Currently over 11,000 in use.

Architecture of Dynamic Routing

AS 1

AS 2

BGP

EGP = Exterior Gateway Protocol

IGP = Interior Gateway Protocol

Metric based: OSPF, IS-IS, RIP, EIGRP (cisco)

Policy based: BGP

The Routing Domain of BGP is the entire Internet

OSPF

EIGRP

• Topology information is flooded within the routing domain

• Best end-to-end paths are computed locally at each router.

• Best end-to-end paths determine next-hops.

• Based on minimizing some notion of distance

• Works only if policy is shared and uniform

• Examples: OSPF, IS-IS

• Each router knows little about network topology

• Only best next-hops are chosen by each router for each destination network.

• Best end-to-end paths result from composition of all next-hop choices

• Does not require any notion of distance

• Does not require uniform policies at all routers

• Examples: RIP, BGP

Link State Vectoring

Technology of Distributed Routing

The Gang of Four

Link State Vectoring

EGP

IGP

BGP

RIPIS-IS

OSPF

20

Many Routing Processes Can Run on a Single Router

Forwarding Table

OSPFDomain

RIPDomain

BGP

OS kernel

OSPF Process

OSPF Routing tables

RIP Process

RIP Routing tables

BGP Process

BGP Routing tables

Forwarding Table Manager

21

IPv4 Addresses are 32 Bit Values

11111111 00010001 10000111 00000000

255 013417

255.17.134.0Dotted quad notation

IPv6 addresses have 128 bits

22

Classful Addresses

0nnnnnnn

10nnnnnn nnnnnnnn

nnnnnnnn nnnnnnnn110nnnnn

hhhhhhhh hhhhhhhh hhhhhhhh

hhhhhhhh hhhhhhhh

hhhhhhhh

n = network address bit h = host identifier bit

Class A

Class C

Class B

Leads to a rigid, flat, inefficient use of address space …

23

RFC 1519: Classless Inter-Domain Routing (CIDR)

IP Address : 12.4.0.0 IP Mask: 255.254.0.0

00001100 00000100 00000000 00000000

11111111 11111110 00000000 00000000

Address

Mask

for hosts Network Prefix

Use two 32 bit numbers to represent a network. Network number = IP address + Mask

Usually written as 12.4.0.0/15

Which IP Addresses are Covered by a Prefix?

00001100 00000100 00000000 00000000

11111111 11111110 00000000 0000000012.4.0.0/15

00001100 00000101 00001001 00010000

00001100 00000111 00001001 00010000

12.5.9.16

12.7.9.16

12.5.9.16 is covered by prefix 12.4.0.0/15

12.7.9.16 is not covered by prefix 12.4.0.0/15

25

CIDR = Hierarchy in Addressing

12.0.0.0/8

12.0.0.0/16

12.254.0.0/16

12.1.0.0/16

12.2.0.0/16

12.3.0.0/16

:::

12.253.0.0/16

12.3.0.0/2412.3.1.0/24

::

12.3.254.0/24

12.253.0.0/1912.253.32.0/1912.253.64.0/19

12.253.96.0/1912.253.128.0/1912.253.160.0/1912.253.192.0/19

:::

Classless Forwarding

Destination =12.5.9.16------------------------------- payload

Prefix Interface Next Hop

12.0.0.0/8 10.14.22.19 ATM 5/0/8

12.4.0.0/15

12.5.8.0/23 attached

Ethernet 0/1/3

Serial 1/0/7

10.1.3.77

IP Forwarding Table

0.0.0.0/0 10.14.11.33 ATM 5/0/9

even better

OK

better

best!

IP Address Allocation and Assignment: Internet Registries

IANAwww.iana.org

RFC 2050 - Internet Registry IP Allocation Guidelines RFC 1918 - Address Allocation for Private Internets RFC 1518 - An Architecture for IP Address Allocation with CIDR

ARINwww.arin.org

APNICwww.apnic.org

RIPEwww.ripe.org

Allocate to National and local registries and ISPs Addresses assigned to customers by ISPs

28

Nontransit vs. Transit ASes

ISP 1ISP 2

Nontransit ASmight be a corporateor campus network.Could be a “content provider”

NET ATraffic NEVER flows from ISP 1through NET A to ISP 2(At least not intentionally!)

IP traffic

Internet Serviceproviders (often)have transit networks

29

Selective Transit

NET BNET C

NET A provides transitbetween NET B and NET Cand between NET D and NET C

NET A

NET D

NET A DOES NOTprovide transitBetween NET D and NET B

Most transit networks transit in a selective manner…

IP traffic

Customers and Providers

Customer pays provider for access to the Internet

provider

customer

IP trafficprovider customer

Customers Don’t Always Need BGP

provider

customer

Nail up default routes 0.0.0.0/0pointing to provider.

Nail up routes 192.0.2.0/24pointing to customer

192.0.2.0/24

Static routing is the most common way of connecting anautonomous routing domain to the Internet. This helps explain why BGP is a mystery to many …

Customer-Provider Hierarchy

IP trafficprovider customer

The Peering Relationship

peer peer

customerprovider

Peers provide transit between their respective customers

Peers do not provide transit between peers

Peers (often) do not exchange $$$trafficallowed

traffic NOTallowed

Peering Provides Shortcuts

Peering also allows connectivity betweenthe customers of “Tier 1” providers.

peer peer

customerprovider

Peering Wars

• Reduces upstream transit costs

• Can increase end-to-end performance

• May be the only way to connect your customers to some part of the Internet (“Tier 1”)

• You would rather have customers

• Peers are usually your competition

• Peering relationships may require periodic renegotiation

Peering struggles are by far the most contentious issues in the ISP world!

Peering agreements are often confidential.

Peer Don’t Peer

36

BGP-4• BGP = Border Gateway Protocol

• Is a Policy-Based routing protocol

• Is the de facto EGP of today’s global Internet

• Relatively simple protocol, but configuration is complex and the

entire world can see, and be impacted by, your mistakes.

• 1989 : BGP-1 [RFC 1105]– Replacement for EGP (1984, RFC 904)

• 1990 : BGP-2 [RFC 1163]

• 1991 : BGP-3 [RFC 1267]

• 1995 : BGP-4 [RFC 1771] – Support for Classless Interdomain Routing (CIDR)

37

BGP Operations (Simplified)

Establish session on TCP port 179

Exchange all active routes

Exchange incremental updates

AS1

AS2

While connection is ALIVE exchangeroute UPDATE messages

BGP session

38

Four Types of BGP Messages

• Open : Establish a peering session.

• Keep Alive : Handshake at regular intervals.

• Notification : Shuts down a peering session.

• Update : Announcing new routes or withdrawing previously announced routes.

announcement = prefix + attributes values

BGP Attributes

Value Code Reference----- --------------------------------- --------- 1 ORIGIN [RFC1771] 2 AS_PATH [RFC1771] 3 NEXT_HOP [RFC1771] 4 MULTI_EXIT_DISC [RFC1771] 5 LOCAL_PREF [RFC1771] 6 ATOMIC_AGGREGATE [RFC1771] 7 AGGREGATOR [RFC1771] 8 COMMUNITY [RFC1997] 9 ORIGINATOR_ID [RFC2796] 10 CLUSTER_LIST [RFC2796] 11 DPA [Chen] 12 ADVERTISER [RFC1863] 13 RCID_PATH / CLUSTER_ID [RFC1863] 14 MP_REACH_NLRI [RFC2283] 15 MP_UNREACH_NLRI [RFC2283] 16 EXTENDED COMMUNITIES [Rosen] ... 255 reserved for development

From IANA: http://www.iana.org/assignments/bgp-parameters

This tutorial will cover these attributes

Not all attributesneed to be present inevery announcement

Attributes are Used to Select Best Routes

192.0.2.0/24pick me!

192.0.2.0/24pick me!

192.0.2.0/24pick me!

192.0.2.0/24pick me!

Given multipleroutes to the sameprefix, a BGP speakermust pick at mostone best route

(Note: it could reject them all!)

41

Two Types of BGP Neighbor Relationships

• External Neighbor (eBGP) in a different Autonomous Systems

• Internal Neighbor (iBGP) in the same Autonomous System AS1

AS2

eBGP

iBGP

iBGP is routed (using IGP!)

42

iBGP Peers Must be Fully Meshed

iBGP neighbors do not announce routes received via iBGP to other iBGPneighbors.

eBGP update

iBGP updates

• iBGP is needed to avoid routing loops within an AS

• Injecting external routes into IGP does not scale and causes BGP policy information to be lost

• BGP does not provide “shortest path” routing

• Is iBGP an IGP? NO!

43

BGP Next Hop Attribute

Every time a route announcement crosses an AS boundary, the Next Hop attribute is changed to the IP address of the border router that announced the route.

AS 6431AT&T Research

135.207.0.0/16Next Hop = 12.125.133.90

AS 7018AT&T

AS 12654RIPE NCCRIS project

12.125.133.90

135.207.0.0/16Next Hop = 12.127.0.121

12.127.0.121

Forwarding Table

Forwarding Table

Join EGP with IGP For Connectivity

AS 1 AS 2192.0.2.1

135.207.0.0/16

10.10.10.10

EGP

192.0.2.1135.207.0.0/16

destination next hop

10.10.10.10192.0.2.0/30

destination next hop

135.207.0.0/16Next Hop = 192.0.2.1

192.0.2.0/30

135.207.0.0/16

destination next hop

10.10.10.10

+

192.0.2.0/30 10.10.10.10

Forwarding Table

Forwarding Table

Next Hop Often Rewritten to Loopback

AS 1 AS 2192.0.2.1

135.207.0.0/16

10.10.10.10

EGP

127.22.33.44135.207.0.0/16

destination next hop

10.10.10.10127.22.33.44

destination next hop

135.207.0.0/16

destination next hop

10.10.10.10

+

127.22.33.44

10.10.10.10

127.22.33.44

135.207.0.0/16Next Hop = 192.0.2.1

135.207.0.0/16Next Hop = 127.22.33.44

Implementing Customer/Provider and Peer/Peer relationships

• Enforce transit relationships – Outbound route filtering

• Enforce order of route preference– provider < peer < customer

Two parts:

Import Routes

Frompeer

Frompeer

Fromprovider

Fromprovider

From customer

From customer

provider route customer routepeer route ISP route

Export Routes

Topeer

Topeer

Tocustomer

Tocustomer

Toprovider

From provider

provider route customer routepeer route ISP route

filtersblock

49

How Can Routes be Colored?BGP Communities!

A community value is 32 bits

By convention, first 16 bits is ASN indicating who is giving itan interpretation

communitynumber

Very powerful BECAUSE it has no (predefined) meaning

Community Attribute = a list of community values.(So one route can belong to multiple communities)

RFC 1997 (August 1996)

Used for signallywithin and betweenASes

Two reserved communities

no_advertise 0xFFFFFF02: don’t pass to BGP neighbors

no_export = 0xFFFFFF01: don’t export out of AS

Communities Example

• 1:100– Customer routes

• 1:200– Peer routes

• 1:300– Provider Routes

• To Customers– 1:100, 1:200, 1:300

• To Peers– 1:100

• To Providers– 1:100

AS 1

Import Export

192.0.2.0/24

192.0.2.0/24Accidental or maliciousannouncement of your prefixcan blackhole your destinations in large part of the Internet

peer peer

customerprovider

Need Filter Here!

legitimate

not legitimate

Blackholes

Mars Attacks!

• 0.0.0.0/0: default• 10.0.0.0/8: private• 172.16.0.0/12: private• 192.168.0.0/16: private• 127.0.0.0/8: loopbacks• 128.0.0.0/16: IANA reserved • 192.0.2.0/24: test networks• 224.0.0.0/3: classes D and E• …..

Martian list often includes

Import Routes (Revisited)

Frompeer

Frompeer

Fromprovider

Fromprovider

From customer

From customer

provider route customer routepeer route ISP route

filtersmartians

xxxxxx

xxxxxx

xxxxxx

xxxxxxxxxxxx

xxxxxxxxxxxx

Customeraddress filters

cccccccccccccccccc

potentialbackhole

Martian

54

So Many Choices

Which route shouldFrank pick to 13.13.0.0./16?

AS 1

AS 2

AS 4

AS 3

13.13.0.0/16

Frank’s Internet Barn

peer peer

customerprovider

55

BGP Route Processing

Best Route Selection

Apply Import Policies

Best Route Table

Apply Export Policies

Install forwardingEntries for bestRoutes.

ReceiveBGPUpdates

BestRoutes

TransmitBGP Updates

Apply Policy =filter routes & tweak attributes

Based onAttributeValues

IP Forwarding Table

Apply Policy =filter routes & tweak attributes

Open ended programming.Constrained only by vendor configuration language

Tweak Tweak Tweak

• For inbound traffic– Filter outbound routes– Tweak attributes on

outbound routes in the hope of influencing your neighbor’s best route selection

• For outbound traffic– Filter inbound routes– Tweak attributes on

inbound routes to influence best route selection

outboundroutes

inboundroutes

inboundtraffic

outboundtraffic

In general, an AS has morecontrol over outbound traffic

Route Selection Summary

Highest Local Preference

Shortest ASPATH

Lowest MED

i-BGP < e-BGP

Lowest IGP cost to BGP egress

Lowest router ID

traffic engineering

Enforce relationships

Throw up hands andbreak ties

58

Back to Frank …

AS 1AS 2

AS 4

AS 3

13.13.0.0/16

peer peer

customerprovider

local pref = 80

local pref = 100

local pref = 90

Higher Localpreference valuesare more preferred

Local preference only used in iBGP

59

Implementing Backup Links with Local Preference (Outbound Traffic)

Forces outbound traffic to take primary link, unless link is down.

AS 1

primary link backup link

Set Local Pref = 100for all routes from AS 1 AS 65000

Set Local Pref = 50for all routes from AS 1

We’ll talk about inbound traffic soon …

60

Multihomed Backups (Outbound Traffic)

Forces outbound traffic to take primary link, unless link is down.

AS 1

primary link backup link

Set Local Pref = 100for all routes from AS 1

AS 2

Set Local Pref = 50for all routes from AS 3

AS 3provider provider

61

ASPATH Attribute

AS7018135.207.0.0/16AS Path = 6341

AS 1239Sprint

AS 1755Ebone

AT&T

AS 3549Global Crossing

135.207.0.0/16AS Path = 7018 6341

135.207.0.0/16AS Path = 3549 7018 6341

AS 6341

135.207.0.0/16

AT&T Research

Prefix Originated

AS 12654RIPE NCCRIS project

AS 1129Global Access

135.207.0.0/16AS Path = 7018 6341

135.207.0.0/16AS Path = 1239 7018 6341

135.207.0.0/16AS Path = 1755 1239 7018 6341

135.207.0.0/16AS Path = 1129 1755 1239 7018 6341

62

Interdomain Loop Prevention

BGP at AS YYY will never accept a route with ASPATH containing YYY.

AS 7018

12.22.0.0/16ASPATH = 1 333 7018 877

Don’t Accept!

AS 1

Traffic Often Follows ASPATH

AS 4AS 3AS 2AS 1135.207.0.0/16

135.207.0.0/16ASPATH = 3 2 1

IP Packet Dest =135.207.44.66

… But It Might Not

AS 4AS 3AS 2AS 1135.207.0.0/16

135.207.0.0/16ASPATH = 3 2 1

IP Packet Dest =135.207.44.66

AS 5

135.207.44.0/25ASPATH = 5

135.207.44.0/25

AS 2 filters allsubnets with maskslonger than /24

135.207.0.0/16ASPATH = 1

From AS 4, it may look like thispacket will take path 3 2 1, but it actually takespath 3 2 5

In fairness: could you do this “right” and still scale?

Exporting internalstate would dramatically increase global instability and amount of routingstate

Shorter Doesn’t Always Mean Shorter

AS 4

AS 3

AS 2

AS 1

Mr. BGP says that path 4 1 is better than path 3 2 1

Duh!

66

Shedding Inbound Traffic with ASPATH Padding Hack

Padding will (usually) force inbound traffic from AS 1to take primary link

AS 1

192.0.2.0/24ASPATH = 2 2 2

customerAS 2

provider

192.0.2.0/24

backupprimary

192.0.2.0/24ASPATH = 2

67

Padding May Not Shut Off All Traffic

AS 1

192.0.2.0/24ASPATH = 2 2 2 2 2 2 2 2 2 2 2 2 2 2

customerAS 2

provider

192.0.2.0/24

192.0.2.0/24ASPATH = 2

AS 3provider

AS 3 will sendtraffic on “backup”link because it prefers customer routes and localpreference is considered before ASPATH length!

Padding in this way is oftenused as a form of loadbalancing

backupprimary

68

COMMUNITY Attribute to the Rescue!

AS 1

customerAS 2

provider

192.0.2.0/24

192.0.2.0/24ASPATH = 2

AS 3provider

backupprimary

192.0.2.0/24ASPATH = 2 COMMUNITY = 3:70

Customer import policy at AS 3:If 3:90 in COMMUNITY then set local preference to 90If 3:80 in COMMUNITY then set local preference to 80If 3:70 in COMMUNITY then set local preference to 70

AS 3: normal customer local pref is 100,peer local pref is 90

69

Hot Potato Routing: Go for the Closest Egress Point

192.44.78.0/24

15 56 IGP distances

egress 1 egress 2

This Router has two BGP routes to 192.44.78.0/24.

Hot potato: get traffic off of your network as Soon as possible. Go for egress 1!

70

Getting Burned by the Hot Potato

15 56

172865High bandwidth

Provider backbone

Low bandwidthcustomer backbone

Heavy Content Web Farm

Many customers want their provider to carry the bits!

tiny http request

huge http reply

SFF NYC

San Diego

71

Cold Potato Routing with MEDs(Multi-Exit Discriminator Attribute)

15 56

172865

Heavy Content Web Farm

192.44.78.0/24

192.44.78.0/24MED = 15

192.44.78.0/24MED = 56

This means that MEDs must be considered BEFOREIGP distance!

Prefer lower MED values

Note1 : some providers will not listen to MEDs

Note2 : MEDs need not be tied to IGP distance

Route Selection Summary

Highest Local Preference

Shortest ASPATH

Lowest MED

i-BGP < e-BGP

Lowest IGP cost to BGP egress

Lowest router ID

traffic engineering

Enforce relationships

Throw up hands andbreak ties

This is somewhat simplified. Hey, what happened to ORIGIN??

Policies Can Interact Strangely(“Route Pinning” Example)

backup

Disaster strikes primary linkand the backup takes over

Primary link is restored but sometraffic remains pinned to backup

1 2

3 4

Install backup link using community

customer

News At 11:00

• BGP is not guaranteed to converge on a stable routing. Policy interactions could lead to “livelock” protocol oscillations. See “Persistent Route Oscillations in Inter-domain Routing” by K. Varadhan,

R. Govindan, and D. Estrin. ISI report, 1996 • Corollary: BGP is not guaranteed to recover

from network failures.

What Problem is BGP solving?

Underlying problem

Shortest Paths

Distributed means of computing a solution.

X?

RIP, OSPF, IS-IS

BGP

• aid in the design of policy analysis algorithms and heuristics• aid in the analysis and design of BGP and extensions• help explain some BGP routing anomalies • provide a fun way of thinking about the protocol

X could

A Wee Bit of Theory

Separate dynamic and static semantics

SPVP = Simple Path Vector Protocol = a distributed algorithm for solving SPP

BGP

SPVP

Booo Hooo, Many, many complications...

BGP Policies

Stable Paths Problem (SPP)

staticsemantics

dynamicsemantics

See [Griffin, Shepherd, Wilfong]

1

An instance of the Stable Paths Problem (SPP)

2 5 5 2 1 0

0

2 1 02 0

1 3 01 0

3 0

4 2 04 3 0

3

4

2

1

• A graph of nodes and edges, • Node 0, called the origin, • For each non-zero node, a set

or permitted paths to the origin. This set always contains the “null path”.

• A ranking of permitted paths at each node. Null path is always least preferred. (Not shown in diagram)

When modeling BGP : nodes represent BGP speaking routers, and 0 represents a node originating some address block

most preferred…least preferred (not null)

Yes, the translation gets messy!

5 5 2 1 0

1

A Solution to a Stable Paths Problem

2

0

2 1 02 0

1 3 01 0

3 0

4 2 04 3 0

3

4

2

1

• node u’s assigned path is either the null path or is a path uwP, where wP is assigned to node w and {u,w} is an edge in the graph,

• each node is assigned the highest ranked path among those consistent with the paths assigned to its neighbors.

A Solution need not represent a shortest path tree, or a spanning tree.

A solution is an assignment of permitted paths to each node such that

An SPP may have multiple solutions

First solution

1

0

2

1 2 01 0

1

0

2

1

0

2

2 1 02 0

1 2 01 0

2 1 02 0

1 2 01 0

2 1 02 0

Second solutionDISAGREE

BAD GADGET : No Solution

2

0

31

2 1 02 0

1 3 01 0

3 2 03 0

4

3

SURPRISE : Beware of Backup Policies

2

0

31

2 1 02 0

1 3 01 0

3 4 2 03 0

4

4 04 2 04 3 0

Becomes a BAD GADGET if link (4, 0) goes down.

BGP is not robust : it is not guaranteed to recover from network failures.

PRECARIOUS

1

0

2

1 2 01 0

2 1 02 0

3

4

5 6

5 3 1 05 6 3 1 2 05 3 1 2 0

6 3 1 06 4 3 1 2 06 3 1 2 0

4 3 1 04 5 3 1 2 04 3 1 2 0

3 1 03 1 2 0

As with DISAGREE, this part has two distinct solutions

This part has a solution only when node 1 is assigned the direct path (1 0).

Has a solution, but can get “trapped”

Part II

Issues of scale for BGP in the real world

Big and Getting Bigger

• Scaling the iBGP mesh– Confederations– Route Reflectors

• BGP Table Growth– Address aggregation (CIDR) – Address allocation

• AS number allocation and use• Dynamics of BGP

– Inherent vs. accidental oscillation – Rate limiting and route flap dampening– Lots and lots of noise – Slow convergence time

ScaleScaleScaleScaleScaleScaleScaleScaleScaleScaleScaleScaleScale

85

iBGP Mesh Does Not Scale

eBGP update

iBGP updates

• N border routers means N(N-1)/2

peering sessions

• Each router must have N-1 iBGP

sessions configured

• The addition a single iBGP speaker

requires configuration changes to all

other iBGP speakers

• Size of iBGP routing table can be order

N larger than number of best routes

(remember alternate routes!)

• Each router has to listen to update

noise from each neighbor

Currently four solutions: (0) Buy bigger routers!(1) Break AS into smaller ASes(2) BGP Route reflectors(3) BGP confederations

86

• Route reflectors can pass on

iBGP updates to clients

• Each RR passes along ONLY

best routes • ORIGINATOR_ID and

CLUSTER_LIST attributes are needed to avoid loops

RR RR

RR

Route Reflectors

BGP Confederations

AS 65501

AS 65502

AS 65503 AS 65504AS 65500

AS 1

From the outside, this looks like AS 1

Confederation eBGP (between member ASes) preserves LOCAL_PREF, MED, and BGP NEXTHOP.

BGP Table Growth

Thanks to Geoff Huston. http://www.telstra.net/ops/bgptable.html on August 8, 2001

Large BGP Tables Considered Harmful

• Routing tables must store best routes and alternate routes

• Burden can be large for routers with many alternate routes (route reflectors for example)

• Routers have been known to die• Increases CPU load, especially during

session reset

Moore’s Law may save us in theory. But in practice it means spending money to upgradeequipment …

Deaggregation Due to Multihoming May be a Leading

Cause

AS 1

customerAS 2

provider

12.0.0.0/8

AS 3provider

12.2.0.0/16

12.2.0.0/16

12.2.0.0/16

If AS 1 doesnot announce themore specific prefix,then most traffic to AS 2 will go through AS 3 because it is a longer match

AS 2 is “punching a hole” inThe CIDR block of AS 1

How Many ASNs are there?

Thanks to Geoff Huston. http://www.telstra.net/ops on June 23, 2001

64,511

2005?2007?

When will we run out of ASNs?

What is to be done?

• Make ASNs larger than 16 bits– How about 32 bits? – See Internet Draft: “BGP support for four-octet AS

number space” (draft-ietf-idr-as4bytes-03.txt)– Requires protocol change and wide deployment

• Change the way ASNs are used– Allow multihomed, non-transit networks to use private

ASNs– Uses ASE (AS number Substitution on Egress )– See Internet Draft: “Autonomous System Number

Substitution on Egress” (draft-jhaas-ase-00.txt)– Works at edge, requires protocol change (for loop

prevention)– Makes some kinds of debugging harder!

Multihomed and “Private”! (draft-jhaas-ase-00.txt)

In fairness: could you “do thisright” and still scale?

AS 2AS 1

AS 65535

63.63.63.0/24

AS 3

63.63.63.0/24

ASPATH = 65535

63.63.63.0/24

ASPATH = 65535

Replaceprivate ASN

Choice of “private” ASN requires a bit of additional coordination between providers

A non-transit network

63.63.63.0/24

ASPATH = 2

ASE-ORIGINATOR = 65535

63.63.63.0/24

ASPATH = 1

ASE-ORIGINATOR = 65535

ASE-ORIGINATOR is a new attribute needed for “sender side loop detection” at AS 1 and 2

BGP Routing Tables

• Use “whois” queries to associate an ASN with “owner” (for example, http://www.arin.net/whois/arinwhois.html)

• 7018 = AT&T Worldnet, 701 =Uunet, 3561 = Cable & Wireless, …

• Hey, we can use these paths to draw cool graphs!

show ip bgpBGP table version is 111849680, local router ID is 203.62.248.4Status codes: s suppressed, d damped, h history, * valid, > best, i - internalOrigin codes: i - IGP, e - EGP, ? - incomplete

Network Next Hop Metric LocPrf Weight Path

. . .*>i192.35.25.0 134.159.0.1 50 0 16779 1 701 703 i*>i192.35.29.0 166.49.251.25 50 0 5727 7018 14541 i*>i192.35.35.0 134.159.0.1 50 0 16779 1 701 1744 i*>i192.35.37.0 134.159.0.1 50 0 16779 1 3561 i*>i192.35.39.0 134.159.0.3 50 0 16779 1 701 80 i*>i192.35.44.0 166.49.251.25 50 0 5727 7018 1785 i*>i192.35.48.0 203.62.248.34 55 0 16779 209 7843 225 225 225 225 225 i*>i192.35.49.0 203.62.248.34 55 0 16779 209 7843 225 225 225 225 225 i*>i192.35.50.0 203.62.248.34 55 0 16779 3549 714 714 714 i*>i192.35.51.0/25 203.62.248.34 55 0 16779 3549 14744 14744 14744 14744 14744 14744 14744 14744 i. . .

Thanks to Geoff Huston. http://www.telstra.net/ops on July 6, 2001

AS Graphs Can Be Fun

The subgraph showing all ASes that have more than 100 neighbors in fullgraph of 11,158 nodes. July 6, 2001. Point of view: AT&T route-server

AS Graphs Depend on Point of View

This explains why there is no UUNET (701) Sprint (1239) link on previous slide!

peer peer

customerprovider

54

2

1 3

6

54

2

6

1 3

54 6

1 3

54

2

6

1 32

AS Graphs Do Not Show Topology!

The AS graphmay look like this. Reality may be closer to this…

BGP was designed to throw away information!

BGP Dynamics

• How many updates are flying around the Internet?

• How long Does it take Routes to Change?

The goals of (1) fast convergence (2) minimal updates (3) path redundancy are at odds

Daily Update Count

What is the Sound of One Route Flapping?

A Few Bad Apples …

Thanks to Madanlal Musuvathi for this plot. Data source: RIPE NCC

Typically, 80% ofthe updates are for less than 5% Of the prefixes.

Most prefixes are stable most of the time. On this day, about 83% of the prefixes were not updated.

Percent of BGP table prefixes

Two BGP Mechanisms for Squashing Updates

• Rate limiting on sending updates– Send batch of updates every

MinRouteAdvertisementInterval seconds (+/- random fuzz)

– Default value is 30 seconds– A router can change its mind about best

routes many times within this interval without telling neighbors

• Route Flap Dampening– Punish routes for misbehaving

Effective in dampening oscillations inherent in the vectoring approach

Must be turned onwith configuration

30 Second Bursts

How Long Does BGP Take to Adapt to Changes?

0

10

20

30

40

50

60

70

80

90

100

0 20 40 60 80 100 120 140 160

Seconds Until Convergence

Cumulative Percentage of Events

Tup

Tshort

Tlong

Tdow n

Thanks to Abha Ahuja and Craig Labovitz for this plot.

Two Main Factors in Delayed Convergence

• Rate limiting timer slows everything down

• BGP can explore many alternate paths before giving up or arriving at a new path– No global knowledge in vectoring protocols

Why is Rate Limiting Needed?

Updatesto convergence

MinRouteAdvertisementInterval

0

Timeto convergence

MinRouteAdvertisementInterval

0

SSFNet (www.ssfnet.org) simulations, T. Griffin and B.J. Premore. To appear in ICNP 2001.

Rate limiting dampens some of the oscillation inherent in a vectoring protocol.

Current interval (30 seconds) was picked “out of the blue sky”

108

Route Flap Dampening (RFC 2439)

Routes are given a penalty for changing. If penalty exceeds suppress limit, the route is dampened. When the route is not changing, its penalty decays exponentially. If the penalty goesbelow reuse limit, then it is announced again.

• Can dramatically reduce the number of BGP updates

• Requires additional router resources• Applied on eBGP inbound only

Route Flap Dampening Example

route dampenedfor nearly 1 hour

penalty for each flap = 1000

Q: Why All the Updates?

• Networks come, networks go• There’s always a router rebooting somewhere • Hardware failure, flaky interface cards, backhoes

digging, floods in Houston, …

This is “normal” --- exactly what dynamic routing is designed for…

Q: Why All the Updates?

• Misconfiguration • Route flap dampening not widely used• BGP exploring many alternate paths• Software bugs in implementation of routing protocols• BGP session resets due to congestion or lack of interoperability: BGP

sessions are brittle. One malformed update is enough to reset session and flap 100K routes. (Consequence of incremental approach)

• IGP instability exported by use of MEDs or IGP tie breaker• Sub-optimal vendor implementation choices• Secret sauce routing algorithms attempting fancy-dancy tricks• Weird policy interactions (MED oscillation, BAD GADGETS??)• Gnomes, sprites, and fairies • ….

A: NO ONE REALLY KNOWS …

112

IGP Tie Breaking Can Export Internal Instability to the Whole Wide World

15 56

192.44.78.0/24

AS 4

AS 3AS 2

AS 1

10

FLAP

FLAP

FLAP FLAP192.44.78.0/24ASPATH = 4 2 1

192.44.78.0/24ASPATH = 4 3 1

113

MEDs Can Export Internal Instability

15

172865

Heavy Content Web Farm

192.44.78.0/24

192.44.78.0/24MED = 15

192.44.78.0/24MED = 56 OR 10

56

10

FLAP

FLAP

FLAP

FLAP

FLAPFLAP

Implementation Does Matter!

Thanks to Abha Ahuja and Craig Labovitz for this plot.

stateless withdrawswidely deployed

stateful withdrawswidely deployed

How Long Will Interdomain Routing Continue to Scale?

... the existing interdomain routinginfrastructure is rapidly nearing the end of its useful lifetime. Itappears unlikely that mere tweaks of BGP will stave off fundamentalscaling issues, brought on by growth, multihoming and other causes.

A quote from some recent email:

Is this true or false? How can we tell?

Research required…

Summary

• BGP is a fairly simple protocol …• … but it is not easy to configure• BGP is running on more than 100K

routers (my estimate), making it one of world’s largest and most visible distributed systems

• Global dynamics and scaling principles are still not well understood

Addressing and ASN RFCs

• RFC 1380 IESG Deliberations on Routing and Addressing (1992) • RFC 1517Applicability Statement for the Implementation of Classless Inter-

Domain Routing (CIDR) (1993) • RFC 1518 An Architecture for IP Address Allocation with CIDR (1993) • RFC 1519 Classless Inter-Domain Routing (CIDR) (1993) • RFC 1467 Status of CIDR Deployment in the Intrenet (1983) • RFC 1520 Exchanging Routing Information Across Provider Boundaries in the

CIDR Environment (1993) • RFC 1817 CIDR and Classful routing (1995) • RFC 1918 Address Allocation for Private Internets (1996) • RFC 2008 Implications of Various Address Allocation Policies for Internet Routing

(1996) • RFC 2050 Internet Registry IP Allocation Guidelines (1996) • RFC 2260 Scalable Support for Multi-homed Multi-provider Connectivity (1998) • RFC 2519 A Framework for Inter-Domain Route Aggregation (1999)

• RFC 1930 Guidelines for creation, selection, and registration of an Autonomous System (AS)

• RFC 2270 Using a Dedicated AS for Sites Homed to a Single Provider

Selected BGP RFCs

• IDR : http://www.ietf.org/html.charters/idr-charter.html

• RFC 1771 A Border Gateway Protocol 4 (BGP-4)

• Latest draft rewrite: draft-ietf-idr-bgp4-12.txt

• RFC 1772 Application of the Border Gateway Protocol in the Internet

• RFC 1773 Experience with the BGP-4 protocol

• RFC 1774 BGP-4 Protocol Analysis

• RFC 2796 BGP Route Reflection An alternative to full mesh IBGP

• RFC 3065 Autonomous System Confederations for BGP

• RFC 1997 BGP Communities Attribute

• RFC 1998 An Application of the BGP Community Attribute in Multi-home Routing

• RFC 2439 Route Flap Dampening

Internet Engineering Task Force (IETF)

http://www.ietf.org

Titles of Some Recent Internet Drafts

• Dynamic Capability for BGP-4• Application of Multiprotocol BGP-4 to IPv4 Multicast Routing • Graceful Restart mechanism for BGP• Cooperative Route Filtering Capability for BGP-4• Address Prefix Based Outbound Route Filter for BGP-4• Aspath Based Outbound Route Filter for BGP-4• Architectural Requirements for Inter-Domain Routing in the Internet• BGP support for four-octet AS number space• Autonomous System Number Substitution on Egress• BGP Extended Communities Attribute• Controlling the redistribution of BGP routes• BGP Persistent Route Oscillation Condition• Benchmarking Methodology for Basic BGP Convergence• Terminology for Benchmarking External Routing Convergence

Measurements

BGP is a moving target …

120

Selected Bibliography on Routing

• Internet Routing Architectures. Bassam Halabi. Second edition Cisco Press, 2000

• BGP4: Inter-domain Routing in the Internet. John W. Stewart, III. Addison-Wesley, 1999

• Routing in the Internet. Christian Huitema. 2000

• ISP Survival Guide: Strategies for Running a Competitive ISP. Geoff Huston. Wiley, 1999.

• Interconnection, Peering and Settlements. Geoff Huston. The Internet Protocol Journal. March and June 1999.

BGP Stability and Convergence

• The Impact of Internet Policy and Topology on Delayed Routing Convergence. Craig Labovitz, Abha Ahuja, Roger Wattenhofer, Srinivasan Venkatachary. INFOCOM 2001

• An Experimental Study of BGP Convergence. Craig Labovitz, Abha Ahuja, Abhijit Abose, Farnam Jahanian. SIGCOMM 2000

• Origins of Internet Routing Instability. C. Labovitz, R. Malan, F. Jahanian. INFOCOM 1999

• Internet Routing Instability. Craig Labovitz, G. Robert Malan and Farnam Jahanian. SIGCOMM 1997

Analysis of Interdomain Routing

• Cooperative Association for Internet Data Analysis (CAIDA)– http://www.caida.org/– Tools and analyses promoting the engineering and maintenance of a

robust, scalable global Internet infrastructure

• Internet Performance Measurement and Analysis (IPMA)– http://www.merit.edu/ipma/– Studies the performance of networks and networking protocols in local

and wide-area networks

• National Laboratory for Applied Network Research (NLANR)– http://www.nlanr.net/– Analysis, tools, visualization.

• IRTF Routing Research Group (IRTF-RR)– http://puck.nether.net/irtf-rr/

Internet Route Registries

• Internet Route Registry– http://www.irr.net/

• Routing Policy Specification Language (RPSL)– RFC 2622 Routing Policy Specification Language

(RPSL) – RFC 2650 Using RPSL in Practice

• Internet Route Registry Daemon (IRRd)– http://www.irrd.net/

• RAToolSet – http://www.isi.edu/ra/RAToolSet/

Some BGP Theory

• Persistent Route Oscillations in Inter-Domain Routing. Kannan Varadhan, Ramesh Govindan, and Deborah Estrin. Computer Networks, Jan. 2000. (Also USC Tech Report, Feb. 1996)

– Shows that BGP is not guaranteed to converge• An Architecture for Stable, Analyzable Internet Routing. Ramesh Govindan, Cengiz Alaettinoglu,

George Eddy, David Kessens, Satish Kumar, and WeeSan Lee. IEEE Network Magazine, Jan-Feb 1999.

– Use RPSL to specify policies. Store them in registries. Use registry for conguration generation and analysis.

• An Analysis of BGP Convergence Properties. Timothy G. Griffin, Gordon Wilfong. SIGCOMM 1999

– Model BGP, shows static analysis of divergence in policies is NP complete• Policy Disputes in Path Vector Protocols. Timothy G. Griffin, F. Bruce Shepherd, Gordon

Wilfong. ICNP 1999– Define Stable Paths Problem and develop sufficient condition for “sanity”

• A Safe Path Vector Protocol. Timothy G. Griffin, Gordon Wilfong. INFOCOM 2001– Dynamic solution for SPVP based on histories

• Stable Internet Routing without Global Coordination. Lixin Gao, Jennifer Rexford. SIGMETRICS 2000

– Show that if certain guidelines are followed, then all is well. • Inherently safe backup routing with BGP. Lixin Gao, Timothy G. Griffin, Jennifer Rexford.

INFOCOM 2001– Use SPP to study complex backup policies

Thank You!

• Companion links:

• http://www.research.att.com/~griffin/interdomain.html