CSE 461: Interdomain Routing

TCP Review: Thought Questions 1. Why do wireless networks often do link-layer re-tx? Hint: is loss always correlated with load? Write a server that accepts new TCP connections but never reads data. Then write a client that opens a TCP connection to that server, and writes forever. Why does write() finally stop working? Where is data stored at that point?

Routing: Scalability Concerns Routing burden grows with size of an inter-network Size of routing tables Volume of routing messages CPU time required for routing computation To scale to the size of the Internet, apply: Hierarchical addressing Route aggregation

Advantages: We can now aggregate routing information. 1/nth as many networks as hosts  fewer updates, smaller tables. Local changes don’t cause global updates. Networks 2, 3

2 2.2 2.3 2.4 2.1 Network 1 Network 3 Network 5 Network 4 The original Internet had exactly 1 level of hierarchy: Network Address and Host Address (Class A, B, C…) From the mid-90’s: CIDR allows arbitrary sub-networking. Further improves route aggregation in the Internet core. Network 4 Network 3 Network 5 Network 1 2 2.2 2.1.1 2.3 2.4 2.1 2.1.2 2.1.3

CIDR Example 128.220.192.0/20 128.220.192.0/19 Decimal Binary X and Y routes can be aggregated because they form a bigger contiguous range. Can only aggregate powers of 2  Corporation X (128.220.192.0 -> 128.220.207.255) 128.220.192.0/20 Border gateway Regional network (advertises path to 128.220.192.0 -> 128.220.223.255) Corporation Y 128.220.192.0/19 (128.220.208.0 -> 128.220.223.255) Decimal Binary 128.192 10000000 11000000 128.207 10000000 11001111 128.208 10000000 11010000 128.223 10000000 11011111 128.220.208.0/20

IP Forwarding Revisited Routing table now contains routes to “prefixes” IP address and length indicating what bits are fixed Now need to “search” routing table for longest matching prefix, only at routers Search routing table for the prefix that the destination belongs to, and use that to forward as before There can be multiple matches; take the longest prefix This is the IP forwarding routine used at routers.

Structure of the Internet Many (~30k) Autonomous Systems (ASes) or domains To scale, use hierarchy: separate inter-domain and intra-domain routing IGP (Interior gateway, within an AS) = RIP, OSPF EGP (Exterior gateway, between Ass) = BGP, EGP (obsolete) Large corporation “ Consumer ” ISP Peering point Backbone service provider Peering point “ Consumer ” ISP “ Consumer ” ISP Large corporation Small corporation

Internet Routing Architecture Divided into Autonomous Systems Distinct regions of administrative control Routers/links managed by a single “institution” Service provider, company, university, … Hierarchy of Autonomous Systems Large, tier-1 provider with a nationwide backbone Medium-sized regional provider with smaller backbone Small network run by a single company or university Interaction between Autonomous Systems Internal topology is not shared between ASes … but, neighboring ASes interact to coordinate routing

Inter-Domain Routing AS1 AS2 Border router Border router Border routers summarize and advertise internal routes to external neighbors and vice-versa Border routers apply policy Internal routers can use notion of default routes Core is “default-free”; routers must have a route to all networks in the world AS1 Border router Border router AS2

AS Topology Node: Autonomous System Edge: Two ASes that connect to each other 1 2 3 4 5 6 7

Interdomain Paths Path: 6, 5, 4, 3, 2, 1 Web server Client 4 3 5 2 7 6

Hierarchical Routing May Pay a Price for Path Quality AS path may be longer than shortest AS path Router path may be longer than shortest path 2 AS hops, 8 router hops d s 3 AS hops, 7 router hops

Border Gateway Protocol (BGP-4) Features: Path vector routing Application of policy Operates over reliable transport (TCP) Uses route aggregation (CIDR)

Internet Interdomain Routing: BGP BGP (Border Gateway Protocol): the de facto standard Path Vector protocol: similar to Distance Vector protocol a border gateway sends to a neighbor entire path (i.e., a sequence of ASes) to a destination, e.g., gateway X sends to neighbor N its path to dest. Z: path (X,Z) = X,Y1,Y2,Y3,…,Z if N selects path(X, Z) advertised by X, then: path (N,Z) = N, path (X,Z) Question: what are the implications of path vector? Z N X

Convergence Prefix P In AS X 1 2 X View from here 4 3 Recently, it was realized that BGP convergence can undergo a process analogous to count-to-infinity! AS 4 uses path 4 1 X. A link fails and 1 withdraws 4 1 X. So 4 uses 4 2 1 X, which is soon withdrawn, then 4 3 2 1 X, … Result is many invalid paths can be explored before convergence Prefix P In AS X 1 2 X View from here 4 3

BGP Routing Decision Process route selection policy: rank paths select best path routing cache export policy: which paths to export to which neighbors export path to neighbors

Business Relationships Neighboring ASes have business contracts How much traffic to carry Which destinations to reach How much money to pay Common business relationships Provider Customer E.g., UW is a customer of NTT E.g., MIT is a customer of Level 3 Peer E.g., AT&T is a peer of Sprint

Customer-Provider Relationship Customer needs to be reachable from everyone Provider tells all neighbors how to reach the customer Customer does not want to provide transit service Customer does not let its providers route through it Traffic to the customer Traffic from the customer advertisements d provider traffic provider customer d customer

Multi-Homing: Two or More Providers Motivations for multi-homing Extra reliability, survive single ISP failure Financial leverage through competition Better performance by selecting better path Gaming the 95th-percentile billing model What implication does this have for routing table size? Provider 1 Provider 2

Peer-Peer Relationship Peers exchange traffic between customers AS exports only customer routes to a peer AS exports a peer’s routes only to its customers Often the relationship is settlement-free (i.e., no $$$) Traffic to/from the peer and its customers advertisements peer peer traffic d

Implication of Business Relationship on Policies Route selection (ranking) policy: the typical route selection policy is to prefer customers over peers/providers to reach a destination, i.e., Customer > Peer > Provider Route export policy: since the export of a path to a neighbor is an indication that the AS is willing to transport traffic for the neighbor, an AS may not export routes to all neighbors

Typical Export Policies customer provider peer case 1: routes learned from customer routes learned from a customer are sent to all other neighbors case 2: routes learned from provider case 3: routes learned from peer provider customer peer peer provider customer routes learned from a provider are sent only to customers routes learned from a peer are sent only to customers

Example Export Policy: No-Valley Routing IP traffic A advertises path to C, but not P2 A advertises path to C, but not P1 A learns paths to C, P1, P2 A A advertises to C paths to P1 and P2 C Suppose P1 and P2 are providers of A; A is a provider of C

AS Structure: Tier-1 Providers Has no upstream provider of its own Typically has a national or international backbone UUNET, Sprint, AT&T, Level 3, … Top of the Internet hierarchy of 9-15 ASes Full peer-peer connections between tier-1 providers

AS Structure: Other ASes Tier-2 and Tier-3 providers Provide transit service to downstream customers … but, need at least one provider of their own Typically have national or regional scope E.g., Minnesota Regional Network Includes a few thousand of the Ases Stub ASes Do not provide transit service to others Connect to one or more upstream providers Includes vast majority (e.g., 85-90%) of the ASes

Characteristics of the AS Graph AS graph structure High variability in node degree (“power law”) A few very highly-connected ASes Many ASes have only a few connections 1 All ASes have 1 or more neighbors 0.1 CCDF 0.01 Very few have degree >= 100 0.001 AS degree 1 10 100 1000

www.cidr-report.org November 2008 Core BGP Table Growth 1994 - 2008 www.cidr-report.org November 2008

% of Entries that advertise /24s www.cidr-report.org November 2008

Begging for Aggregation --- 13Nov08 --- ASnum   NetsNow   NetsAggr   NetGain   % Gain   Description   Table   287512   176594   110918   38.6%   All ASes AS4538   5057   874   4183   82.7%   ERX-CERNET-BKB China Education and Research Network Center AS6389   4372   358   4014   91.8%   BELLSOUTH-NET-BLK - BellSouth.net Inc. AS209   3038   1350   1688   55.6%   ASN-QWEST - Qwest Communications Corporation AS6298   2100   729   1371   65.3%   ASN-CXA-PH-6298-CBS - Cox Communications Inc. AS17488   1409   290   1119   79.4%   HATHWAY-NET-AP Hathway IP Over Cable Internet AS4755   1320   239   1081   81.9%   TATACOMM-AS TATA Communications formerly VSNL is Leading ISP AS1785   1695   618   1077   63.5%   AS-PAETEC-NET - PaeTec Communications, Inc. www.cidr-report.org November 2008

Key Concepts Internet is a collection of Autonomous Systems (ASes) Policy dominates routing at the AS level Structural hierarchy helps make routing scalable BGP routes between autonomous systems (ASes)