Computer Networks:

Intra-AS & Inter-AS

Routing Architecture
CS455/CS555/EE407/EE507
Fall 2024

Pat Wilbur
[email protected]
Dynamic routing in the real world

Our study of dynamic routing protocols so far is an idealized graph problem:
➢ All routers identical
➢ Network “flat”
➢ … not true in practice

Internet scale with more than hundreds of million destinations:
➢ Can’t store all destinations in routing tables!
➢ Routing table exchange would swamp links!
➢ … naive versions of both link-state and distance-vector algorithms
wouldn’t be able to handle the entire Internet
*
Dynamic routing in the real world

Administrative autonomy:
➢ Internet = network of networks
➢ Each network controls routing in its own network

Global routing system to route between Autonomous Systems (inter-AS)

Autonomous System is a connected group of IP networks that adhere to a
single unique routing policy that differs from the routing policies of its
network's border peers

Two-level routing/hierarchical routing:
➢ Intra-AS: administrator is responsible for choice of routing algorithm
➢ Inter-AS: common standard to which everyone adheres *
Internet at the “top levels”

Source
Internet hierarchy of ASes

Source
Hierarchical routing

Routers in same AS run routing gateway routers
protocol chosen by administrators
of that domain (intra-AS routing 
Special routers in AS
protocol) 
Run intra-AS routing

Routers in different AS can run protocol with all other
different intra-AS routing routers in AS
protocols 
Also responsible for routing
to destinations outside AS
(also run same inter-AS
routing protocol as with
other gateway routers)

*
Intra-AS and inter-AS routing
C.b
B.a
A.a
b A.c c
a a
AS C
a AS B b
d c
AS A b

Gateway routers:
perform inter-AS routing amongst themselves and
perform intra-AS routers with other routers in their AS

*
Intra-AS and inter-AS routing
Inter-AS border (exterior side) of gateway routers
C.b
B.a
A.a
b A.c c
a a
AS C
a AS B b
d c
AS A b
Intra-AS (interior side) of gateway routers
Gateway routers:
perform inter-AS routing amongst themselves and
perform intra-AS routers with other routers in their AS

*
Intra-AS and inter-AS routing
C.b
B.a
A.a
b A.c c
a a
AS C
a AS B b
d c
AS A b

Gateway routers:
perform inter-AS routing amongst themselves and
perform intra-AS routers with other routers in their AS

*
Intra-AS and inter-AS routing
C.b
B.a
A.a
b A.c c
a a
AS C
a AS B b
d c
AS A b

inter-AS, intra-AS internet/network layer

routing in data/link layer
gateway A.c physical layer

*
Intra-AS and inter-AS routing
C.b
B.a
A.a
b A.c c
a a Host
AS C
a AS B b
h2
Host d c
h1 AS A b

*
Intra-AS and inter-AS routing
C.b
B.a
A.a
b A.c c
a a Host
AS C
a AS B b
h2
Host d c
h1 AS A b

Single packet is often routed over many hops via routes established by
several intra-AS routing protocols and an inter-AS routing protocol

*
Intra-AS and inter-AS routing
C.b
B.a
A.a
b A.c c
a a Host
AS C
a AS B b
h2
Host d c
h1 AS A b
Intra-AS routing within AS A

Single packet is often routed over many hops via routes established by
several intra-AS routing protocols and an inter-AS routing protocol

*
Intra-AS and inter-AS routing
Inter-AS routing between A and B
C.b
B.a
A.a
b A.c c
a a Host
AS C
a AS B b
h2
Host d c
h1 AS A b
Intra-AS routing within AS A

Single packet is often routed over many hops via routes established by
several intra-AS routing protocols and an inter-AS routing protocol

*
Intra-AS and inter-AS routing
Inter-AS routing between A and B
C.b
B.a
A.a
b A.c c
a a Host
AS C
a AS B b
h2
Host d c Intra-AS routing within AS B
h1 AS A b
Intra-AS routing within AS A

Single packet is often routed over many hops via routes established by
several intra-AS routing protocols and an inter-AS routing protocol

*
Intra-AS vs. inter-AS protocols

Intra-AS routing protocols: many choices
Inter-AS routing protocols: standard – why?

Intra-AS routing protocols: focus upon performance optimization
Inter-AS routing protocols: focus upon administrative issues

Intra-AS routing: often static routing based on CIDR, can also be dynamic
(for dynamic, usually RIP protocol or OSPF protocol)
Inter-AS routing: dynamic routing (BGP protocol)

*
Intra-AS routing

Also known as Interior Gateway Protocols (IGPs)

Example IGPs:
➢ RIP: Routing Information Protocol
➢ OSPF: Open Shortest Path First
➢ (E)IGRP: Interior Gateway Routing Protocol (Cisco proprietary)
➢ Static (CIDR entries added statically to route tables)

*
Routing Information Protocol (RIP)

Implementation of distance-vector algorithm

Included in BSD-UNIX Distribution since 1982

Single path distance (path cost) metric: # of hops (max = 15 hops)
➢ Why? Count-to-infinity problem less painful if infinity = 16
➢ But limits RIP to networks with a diameter of 15 hops

Distance vectors exchanged every 30 seconds via Response Message (also
called advertisement)

Each advertisement: route to up to 25 destination (sub)nets

Each link has cost 1, so path cost truly is just # hops
*
RIP: Link failure and recovery

If no advertisement heard after 180 sec → neighbor/link declared dead
➢ routes via neighbor invalidated
➢ new advertisements sent to neighbors
➢ neighbors in turn send out new advertisements (if route tables changed)
➢ link failure info quickly propagates to entire net
➢ small loops are prevented using “poison reverse” mitigation method
(similar to sender header already discussed)
➢ infinite distance = 16 hops → makes count-to-infinity problem with larger
loops less painful

*
RIP: Route table processing

RIP routing tables managed by application-level process called routed
(daemon)

Advertisements sent in UDP packets, periodically repeated

routed process periodically adjusts OS kernel routing table with updates

*
Open Shortest Path First (OSPF)

“open”: publicly available

Implementation of link-state algorithm:
➢ Entire network topology map (graph) stored at each node
➢ Route computation using Dijkstra’s algorithm

OSPF advertisement carries one entry per neighbor router (i.e. link cost to
each neighbor)

Advertisements disseminated to entire AS (via “reliable flooding”) carried in
OSPF protocol messages directly over IP (rather than TCP or UDP)

*
OSPF: Extra features (not in RIP)

Security: all OSPF messages authenticated (to prevent malicious intrusion)

Uses TCP connections (so, acknowledgement)

Multiple same-cost paths can be used at once (single path need not be
chosen as in RIP)

For each link, multiple cost metrics for different “TOS” (types of service):
e.g. high bandwidth, high-delay satellite link cost may set “low” for best effort
or high for real time

Integrated unicast and multicast support: multicast OSPF (MOSPF) uses
same topology data base as OSPF

Hierarchical OSPF in large domains (large ASes)
*
OSPF: Hierarchical OSPF

Two-level hierarchy: local area, backbone area
➢ Link-state advertisements only in level area
➢ Each node has detailed level area topology; only know direction
(shortest path) to nets in other level areas

Area border routers: “summarize” distances to nets in own level area,
advertise to other area border routers

Backbone routers: run OSPF routing limited to backbone area

Boundary routers: connect an AS to other ASes

*
OSPF: Hierarchical OSPF
Within each level area, border router
responsible for routing outside the
level area Exactly one level area is
the backbone area

Backbone area contains all area border routers and possibly others

*
Interior Gateway Routing Protocol

Cisco proprietary, successor of RIP (mid 1980s)

Distance-vector, like RIP but with advanced features similar to OSPF

Several cost metrics (delay, bandwidth, reliability, load, etc) and
administrator decides which cost metrics to use

Uses TCP to exchange routing updates

Loop-free routing via algorithm known as “diffusing update algorithm”
(DUAL finite-state machine) based upon diffused computation

*
Now on to inter-AS routing...
C.b
B.a
A.a
b A.c c
a a
AS C
a AS B b
d c
AS A b

*
Autonomous systems

The Global Internet consists of Autonomous Systems (AS), which are
distinct networks, all interconnected with each other:
➢ Stub AS: small corporation
➢ Multihomed AS: large corporation (no transit traffic)
➢ Transit AS: provider (carries transit traffic)

A major goal of the inter-AS routing protocol is to reduce/minimize transit
traffic (ASes only want to carry traffic for their own customers because of
costs/capitalism/etc.)

Every AS needs an globally unique AS Number (ASN) handed out by
regional registries, used for identifying the AS and exhcanging routing
information with neighboring ASes
*
Inter-AS (Internet) routing: BGP

BGP (Border Gateway Protocol): the de facto standard on the Internet

Implements a “path-vector” algorithm/protocol:
➢ Similar to distance-vector algorithm
➢ Avoids count-to-infinity problem by identifying yourself in a path
advertised to you
➢ Each Border Gateway broadcast to neighbors (peers) includes entire
path (i.e. sequence of ASNs) to destination ASN

e.g., Gateway X may send its path to destination Z:

Path (X→Z) = X,Y1,Y2,Y3,…,Z

*
BGP basics

BGP session: two BGP routers (peers) exchange BGP messages:
➢ Advertising paths to different destination network prefixes (“path-vector”)
➢ Exchanged over TCP connections

e.g. when AS3 advertises a prefix to AS1:
➢ AS3 promises it will forward packets towards that network prefix
➢ AS3 can aggregate prefixes in its advertisement
3c BGP
3a message
3b 2c
AS3 1c other
2a networks
other 1a 2b
networks 1b AS2
AS1 1d *
BGP basics

Using eBGP session between 3a and 1c, AS3 sends prefix reachability info
to AS1
➢ 1c can then use iBGP to distribute new prefix info to all routers in AS1
➢ 1b can then re-advertise new reachability info to AS2 over the 1b-to-2a
eBGP session

When router learns of new prefix, it creates entry for prefix in its forwarding
(route) table
3c eBGP session
3a iBGP session
3b 2c
AS3 1c other
2a networks
other 1a 2b
networks 1b AS2
AS1 1d *
BGP path attributes & routes

Two important attributes:
➢ AS-PATH: contains ASes through which prefix advertisement has
passed (e.g. [AS 67, AS 17, ...])
➢ NEXT-HOP: indicates specific internal-AS router to next-hop AS (can be
multiple links from current AS to next-hop AS)

Gateway router receiving route advertisement uses import policy to
accept/decline:
➢ e.g., never route through AS x
➢ policy-based routing

*
BGP path attributes & routes

Router may learn about more than 1 route to destination AS, selects route
based upon:
➢ Local preference value attribute: policy decision
➢ Shortest AS-PATH
➢ Closest NEXT-HOP router
➢ Additional criteria

*
BGP path attributes & routes

Suppose: Gateway X send its path to peer Gateway W
➢ W may or may not select path offered by X: e.g. cost, policy (don’t route
via competitor’s AS), loop prevention reasons
➢ If W selects the path advertised by X, then:
Path (W,Z) = W, Path (X,Z)
➢ Note: X can control incoming traffic by controlling its route
advertisements to peers: e.g., don’t want to route traffic to Z → don’t
advertise any routes to Z

*
BGP messages

OPEN: opens TCP connection to peer and authenticates sender

UPDATE: advertises new path (or withdraws old path)

KEEPALIVE keeps connection alive in absence of UPDATES, also
acknowledges OPEN request

NOTIFICATION: reports errors in previous message, also used to close
connection

*
BGP routing policy example 1
legend: provider
B network
X
W A
customer
C
network:
Y

 A, B, and C are provider networks

 X, W, and Y are customers (of provider networks)
 X is dual-homed: attached to two networks
 X does not want to route from B via X to C ...
 ... so X will not advertise to B a route to C

*
BGP routing policy example 2
legend: provider
B network
X
W A
customer
C
network:
Y

 A advertises path [A,W] to B

 B advertises path [B,A,W] to X
 Should B advertise path [B,A,W] to C?
 No way! B gets no revenue for routing C via [B,A,W] since neither W
nor C are B’s customers, B wants to route only to/from its customers
 B wants to force C to route to W via A
*
Why different intra-/inter-AS routing?

Policy:
➢ Inter-AS: admin wants control over how its traffic routed, who routes
through its network
➢ Intra-AS: single admin, so no policy decisions needed

Scale: Hierarchical routing saves table size, reduced update traffic

Performance:
➢ Intra-AS: can focus on performance
➢ Inter-AS: policy may dominate over performance

*
References & attribution:
Original slides by Patrick F. Wilbur, 2024.

Other slides adapted from slides by Professor Yu Liu and Professor Jeanna Matthews.

In this file, slides marked * are heavily adapted or taken from slides by Professor Yu Liu and Professor Matthews.

Other references are depicted on individual slides themselves.