Chapter 4: Network Layer

4. 1 Introduction
4.2 Virtual circuit and

datagram networks
4.3 Whats inside a
router
4.4 IP: Internet
Protocol

Datagram format
IPv4 addressing
ICMP
IPv6

4.5 Routing algorithms

Link state
Distance Vector
Hierarchical routing
4.6 Routing in the

Internet

RIP
OSPF
BGP

4.7 Broadcast and

multicast routing

Hierarchical OSPF
Perhaps some routers dont need to know about every link.

E
C
G
A

ABR
ABR
ABR
ABR

two-level hierarchy: local area,

backbone.
Link-state advertisements
only within the area
each nodes has detailed
knowledge of its area
topology
area border routers:
summarize distances to nets
in own area, advertise to other
Area Border routers.
backbone routers: run OSPF
routing limited to backbone.
boundary routers: connect to
other ASs.

C announces link C<->A to Area 1

C announces link C<->E to Backbone
C announces a summary of Area 1 to the Backbone
C announces a summary of the Backbone and other areas to Area 1
C learns about other areas from the other ABR

Area Border Router Summaries

Should the summaries include reachbility information or path metrics?

Routers in area 1 do not need to

C
G

know about the paths used to

reach destinations in other areas
They only need to know that
they can be reached.
In this case, reachbility
information is sufficient to
compute optimal routes
i.e., the ABR only announces
which destinations it can reach.
However, no one would make a
topology as shown in the figure
Why?
If a single key links break or
router crashes, the network
would be partitioned (and the
network designer would be fired)

Area Border Router Summaries

e.g., if summaries only include reachbility information
area border router

C
G
A
B
D
ABR C announces to Area 1 that it can reach Area 2 in 1 hops (and

includes a list of destinations in Area 2)

ABR F announces to Area 1 that it can reach Area 2 in 0 hops
Router A determines the path to D as follows
The path to Area 2 via F is 2 hops (2 to reach F and then 0
more to Area 2)
The path to Area 2 via C is 2 hops (1 to C and then 1 more to
Area 2)
Either path is good to reach D
However, the path via F is better. A does not have sufficient
information to determine this.

Area Border Router Summaries

In this case, reachability
area border router

C
G
A
B
D

ABR G tells all routers in the Backbone that it can reach D in 2 hop.
ABR F tells all routers in the Backbone that it can reach D in 1 hops
ABR C tells all routers in Area 1 that it can reach D in 3 hops
ABR F tells all routers in Area 1 that it can reach D in 1 hop
A decides B is the best next hop toward D

information is not enough to

compute optimal routes.
Therefore, ABRs provide
distance vector type information,
i.e., which destinations can be
reached and the cost to reach
them
Notice the C does not announce the

link CG to Area 1.
Notice that C gets a summary from G,
which is distances to destinations, like
distance vector.
C uses the distances from G to
determine its distances.
C announces these distances to Area 1
This is like a one hop distance vector
protocol

Area Border Router Summaries

area border router

C
G

The backbone is completely connected

because each router essentially sends
distance vector updates directly to its
neighbor

A
B
D

A in 1 hop
B in 2 hops

Area 1

A in 4 hop
B in 5 hops

A in 2 hop
B in 3 hops

Area 3

Area 2

This is like a one hop distance

vector protocol
Convergence time: 1
Loops are not possible

Chapter 4: Network Layer

4. 1 Introduction
4.2 Virtual circuit and

datagram networks
4.3 Whats inside a
router
4.4 IP: Internet
Protocol

Datagram format
IPv4 addressing
ICMP
IPv6

4.5 Routing algorithms

Link state
Distance Vector
Hierarchical routing
4.6 Routing in the

Internet

RIP
OSPF
BGP

4.7 Broadcast and

multicast routing

Recall: Subnets

223.1.1.2

223.1.1.1

223.1.1.4
223.1.1.3

223.1.9.2

223.1.7.0

223.1.9.1

223.1.7.1
223.1.8.1

223.1.8.0

223.1.2.6
223.1.2.1

223.1.3.27
223.1.2.2

223.1.3.1

223.1.3.2

IP addressing: CIDR
CIDR: Classless InterDomain Routing
subnet portion of address of arbitrary length
address format: a.b.c.d/x, where x is # bits in
subnet portion of address

Subnet part or
CIDR-block

host
part

11001000 00010111 00010000 00000000

200.23.16.0/23

IP addresses: how to get one?

Q: How does network get subnet part of IP addr?
A: gets allocated portion of its provider ISPs
address space

ISP's block

11001000 00010111 00010000 00000000

200.23.16.0/20

Organization 0
Organization 1
Organization 2
...

11001000 00010111 00010000 00000000

11001000 00010111 00010010 00000000
11001000 00010111 00010100 00000000
..
.

200.23.16.0/23
200.23.18.0/23
200.23.20.0/23
.

Organization 7

11001000 00010111 00011110 00000000

200.23.30.0/23

Hierarchical addressing: route aggregation

Hierarchical addressing allows efficient advertisement of routing
information:
Organization 0

200.23.16.0/23
Organization 1

200.23.18.0/23

Organization 2

200.23.20.0/23

Organization 7

.
.
.

.
.
.

ISP1

Send me anything
with addresses
beginning
200.23.16.0/20
Border Router

200.23.30.0/23
ISP2

Send me anything
with addresses
beginning
199.31.0.0/16

This way, the whole 32 bit address does not need to be examined

Internet

Hierarchical addressing: more specific

routes
ISP2 has a more specific route to Organization 1
Organization 0

200.23.16.0/23

Organization 2

200.23.20.0/23

Organization 7

.
.
.

.
.
.

ISP1

Send me anything
with addresses
beginning
200.23.16.0/20
Border Router

200.23.30.0/23
ISP2
Organization 1

200.23.18.0/23

Send me anything
with addresses
beginning 199.31.0.0/16
or 200.23.18.0/23

Internet

Longest prefix matching

Border Router Forwarding Table
Prefix Match
200.23.16.0/20
200.23.18.0/23
199.31.0.0/16
otherwise

Link Interface
0
1
1
2

If a packet with destination address 200.23.18.12 arrives at the boarder

router, then is it forwarding to interface 0 or 1?
Since interface 1 has a longer match, it goes to interface 1

A Problem with Longest Match and subnetting

In order to improve reliability, organization 7 has a backup link with ISP1.
This way, if ISP1 has problems or ISP1s provider has problems, then
organization 7 is still reachable.
Will this work?
Organization 0

200.23.16.0/23
Organization 1

200.23.18.0/23

Organization 2

200.23.20.0/23

Organization 7

.
.
.

.
.
.

ISP1

Send me anything
with addresses
beginning

Border Router

200.23.30.0/23
ISP2

Send me anything
with addresses
beginning
..

Internet

Hierarchical Routing
Our routing study thus far has been an idealization
all routers identical
network flat
not true in practice
scale: with 200 million destinations:

cant store all dests in routing tables!

Memory for address table must be very fast

How fast? How long can an address lookup take on a 10GBit

interface?
E.g., 64B/10 10=50nsec

routing table exchange would swamp links!

There are ~ 1 million links

If link state was flooded every 30 minutes seconds

and each link state is 20B, then each router receives
and processes 100kbps in link announcements

But, perhaps, only changes in link state could be

distributed.

administrative autonomy

internet = network of networks

each network admin wants to control
routing in its own network

ATT does not want Sprint to know what

their topology is

Trade secret
Improves security

ATT wants to select a routing protocol and

parameters without getting Sprints
permission

Hierarchical Routing

aggregate routers into regions,

autonomous systems (AS)
Single administrative domain
Routers in the same AS run same
routing protocol

intra-AS routing protocol

routers in different AS can run
different intra-AS routing
protocol

An ISP may be made of 1 or more

ASs

ATT-USA = 1 AS and ATT-Europe

is another
Some stub networks are an AS

UD is an AS
Some companies have routers but
are not Ass

ASs have their own number,

assigned by ICANN
There are ~50K ASs

Gateway router
Direct link to router in another
AS
Gateway routers run a common
inter-networking routing protocol

Simple example
Connections to other ASs and the rest of the Internet

AS2

Service provider of AS1 (e.g., AS1=UD and AS2=cogent)

The rest of the internet

(Recall that ASs (ISPs) sometimes

meet at NAPs. E.g., google: MAE-East)
An AS could also meet its provider at
a POP.

Stub network (at the edge of the network)

These tables are made with
RIP, OSPF, ISIS, etc

Forwarding table
Interface
Prefix
4
10.1.1.0/24
3
10.1.2.0/24
2
10.1.4.0/22

10.1.1.0/24

4 C
3

1
B
2

Forwarding table
Prefix
3
10.1.1.0/24
3
10.1.2.0/24
2
10.1.4.0/22

AS1

3
A
2

10.1.2.0/24
10.1.4.0/22

Forwarding table
Interface
Prefix
3
10.1.1.0/24
3
10.1.2.0/24
2
10.1.4.0/22

Q: How can routers in AS1 know

where to send pkts with
destination not in AS1?
A: Easy, if a pkt is for an unknown
address, then send it to B.
Specifically, B advertises a link
to prefix 0.0.0.0/0
This is called a default route,
and it can be statically set (no
need for any routing protocol
beside OSPF)

AS2

Service provider of AS1 (e.g., AS1=UD and AS2=cogent)

The rest of the internet

Stub network (at the edge of the network)

These tables are made with
RIP, OSPF, ISIS, etc

Forwarding table
Interface
Prefix
4
10.1.1.0/24
3
10.1.2.0/24
2
10.1.4.0/22
1
0.0.0.0/0

10.1.1.0/24

4 C
3

1
B
2

Forwarding table
Prefix
3
10.1.1.0/24
3
10.1.2.0/24
2
10.1.4.0/22
1
0.0.0.0/0

3
A
2

10.1.2.0/24
10.1.4.0/22

Forwarding table
Interface
Prefix
3
10.1.1.0/24
3
10.1.2.0/24
2
10.1.4.0/22
1
0.0.0.0/0

AS1

We need to put prefixes 1.1.0.0/16, 1.2.0.0/16, 2.2.0.0/16 in the forwarding tables

How to get there?
1.
B must learn from E that 1.1.0.0/16 and 1.2.0.0/16 are reachable through E
2.
A must learn that 2.2.0.0/16 is reachable through D
3.
B and A must distribute this information throughout AS1
Steps 1 and 2 need a exterior inter-networking routing protocol
Step 3 needs an interior inter-networking routing protocol
EBGP and IBGP border gateway routing protocol can accomplish this
to the rest of
the Internet
1.2.0.0/16
1.1.0.0/16

AS2

These tables are made with RIP, OSPF, ISIS, etc

Forwarding table
Interface
Prefix
4
10.1.1.0/24
3
10.1.2.0/24
2
10.1.4.0/22

10.1.1.0/24

1
B

4 C 1
2
3
3

10.1.1.1/24
10.1.4.0/22

Forwarding table
Prefix
3
10.1.1.0/24
3
10.1.2.0/24
2
10.1.4.0/22

1
A
2

AS1

Forwarding table
Interface
Prefix
3
10.1.1.0/24
3
10.1.2.0/24
2
10.1.4.0/22

AS3

2.2.0.0/16

Interconnected ASes
3c

3a
3b
AS3
1a

2a

1c
1d

1b

Intra-AS
Routing
algorithm

2c
AS2

AS1

Inter-AS
Routing
algorithm

Forwarding
table

2b

forwarding table

configured by both
intra- and inter-AS
routing algorithm

intra-AS sets entries

for internal dests
inter-AS & intra-As
sets entries for
external dests

Example: Setting forwarding table in router 1d

suppose AS1 learns (via inter-AS protocol) that subnet x reachable via AS3
(gateway 1c) but not via AS2.
inter-AS protocol propagates reachability info to all internal routers.
router 1d determines from intra-AS routing info that its interface I is on the least
cost path to 1c.
installs forwarding table entry (x,I)
Alternatively, 1d has two table entries

One entry says x is reachable via 1c (determined by IBGP)

A second entry says which is the next hop to reach 1c (determined by intra-routing protocol)

3c

3a
3b
AS3

1a

x
2a

1c
1d

1b AS1

2c

2b
AS2

Example: Choosing among multiple ASes

now suppose AS1 learns from inter-AS protocol that subnet

is reachable from AS3 and from AS2.

to configure forwarding table, router 1d must determine
towards which gateway it should forward packets for dest x.
this is also job of inter-AS routing protocol!
If both gateways are equivalent, then the intra-AS routing
protocol will route packets to the best gateway
This is called hot potato routing: send packet towards closest of two
routers.

3c

3a
3b
AS3

1a

2a

1c
1d

1b

2c
AS2

AS1

2b

Hot Potato Routing

128.4.0.0/16

AS1

AS2

Pkt arrives with dest in 124.4.0.0/16

AS2 could give send the pkt to gateway B hot potato routing.
But AS1 would prefer AS2 to carry its own traffic.
So AS1 might require that AS2 gives higher priority to gateway A.
But how can AS1 enforce AS2 to do this?

Example: Choosing among multiple ASes

now suppose AS1 learns from inter-AS protocol that

subnet x is reachable from AS3 and from AS2.

to configure forwarding table, router 1d must
determine which gateway it should forward packets
for dest x.
this is also job of inter-AS routing protocol!
hot potato routing: send packet towards closest of
two routers.
Learn from inter-AS
protocol that subnet
x is reachable via
multiple gateways

Use routing info

from intra-AS
protocol to
determine
costs of least-cost
paths to each
of the gateways

Hot potato routing:

Choose the
gateway
that has the
least cost

Determine from
forwarding table the
Interface that leads
to least-cost gateway.
Enter (x,I) in
forwarding table

Internet inter-AS routing: BGP

BGP (Border Gateway Protocol):

the de

facto standard
BGP provides each AS a means to:
1.
2.
3.

Obtain subnet reachability information from

neighboring ASs.
Propagate reachability information to all ASinternal routers.
Determine good routes to subnets based on
reachability information and policy.

allows subnet to advertise its existence to

rest of Internet: I am here

BGP basics
pairs of routers (BGP peers) exchange routing info over semi-

permanent TCP connections: BGP sessions

BGP sessions need not correspond to physical links.
when AS2 advertises a prefix to AS1:
AS2 promises it will forward datagrams towards that prefix.
AS2 can aggregate prefixes in its advertisement
But this can cause problems when some prefixes have backup links

eBGP session

3c
3a
3b
AS3
1a
AS1

iBGP session

2a

1c
1d

1b

2c
AS2

2b

Distributing reachability info

using eBGP session between 3a and 1c, AS3 sends prefix

reachability info to AS1.

1c can then use iBGP do distribute new prefix info to all
routers in AS1
1b can then re-advertise new reachability info to AS2 over 1bto-2a eBGP session
when router learns of new prefix, it creates entry for prefix in
its forwarding table.

eBGP session

3c
3a
3b
AS3
1a
AS1

iBGP session

2a

1c
1d

1b

2c
AS2

2b

Aggregation Problem
1.1.1.0/24
1.1.2.0/24

1.1.0.0/22

ISP

ISP
Y

1.1.3.0/24

1.1.4.0/24

1.1.0.0/22
ISP
W

1.1.4.0/24

ISP

1.1.4.0/24

ISP

ISP

ISP
X

From ISP W, the next hop to 1.1.4.0/24 is X, it should be Y

Path attributes & BGP routes

advertised prefix includes BGP attributes.
prefix + attributes = route
two important attributes:

AS-PATH: contains ASs through which prefix

advertisement has passed: e.g, AS 67, AS 17,
NEXT-HOP: indicates specific internal-AS router to
next-hop AS. (there may be multiple routers with
links from current AS to next-hop-AS. Each router
can advertise the path)

when gateway router receives route

advertisement, uses import policy to

accept/decline.

BGP route selection

router may learn about more than 1 route
to some prefix. Router must select route.
elimination rules:

1.
2.
3.
4.

local preference value attribute: policy

decision
shortest AS-PATH
closest NEXT-HOP router: hot potato routing
additional criteria

BGP messages
BGP messages exchanged using TCP.
BGP messages:

OPEN: opens TCP connection to peer and

authenticates sender
UPDATE: advertises new path (or withdraws old)
KEEPALIVE keeps connection alive in absence of
UPDATES; also ACKs OPEN request
NOTIFICATION: reports errors in previous msg;
also used to close connection

BGP routing policy

legend:

B
W

provider
network
customer
network:

C
Y

A,B,C are provider networks

X,W,Y are customer (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 (2)

legend:

B
W

provider
network
customer
network:

C
Y

A advertises path AW to B
B advertises path BAW to X
Should B advertise path BAW to C?

No way! B gets no revenue for routing CBAW since

neither W nor C are Bs customers
B wants to force C to route to w via A
B wants to route only to/from its customers!

BGP route processing

BGP advertises and withdraws paths with the UPDATE message

UPDATE has three fields

Router to withdraw
Attributes of routes to prefixes in NLRI
NLRI

The NLRI is a list of prefixes that the list of attributes applies to. If two prefixes have
different attributes, then these two prefixes need to be announced with different
UPDATE messages.
In OSPF each path is a list of routes and a total cost (two attributes). In BGP, routes have
many attributes, the cost (in AS hops) is only one of the attributes

configuration
from
peers

input
policy routing
engine decision

configuration
routing
table

output
policy
engine

to
peers

RIBs

Routing information base (RIB) a list of routes (including attributes)

Adj-RIB-In: RIB learned from neighbor (many of these)

Adj-RIB-Out: RIB to be sent to neighbor (many of these)
Loc-RIB: RIB for local use (only one of these)

peer

Adj-rib-in

Adj-rib-out

peer

Adj-rib-in

Adj-rib-out

peer

Adj-rib-out

peer

Adj-rib-out

peer

peer

peer

Adj-rib-in

Adj-rib-in

Input
Policy
engine

BGP
Loc-RIB
decision

Input
Policy
engine

peer

Sample routing environment

deny 0/0 from AS1
Give 192.213.1.0/24 from

AS1

Do not propagate 0/0

Do not send 193.214.10.0/24 to AS4
Give 192.213.1.0/24 with metric = 10

AS1 better preference

Accept other routes

AS3

to AS3

192.213.1.0/24
0/0

AS2

193.214.10.0/24 path=(AS5, AS2)

192.213.1.0/24 path=(AS5, AS1)

input
policy
engine

192.213.1.0/24
193.214.10.0/24
0/0

decision
process

routes

Use
Use
Use
Use

0/0 from AS2

192.213.1.0/24 from AS1
193.214.10.0/24 from AS2
172.16.10.0/24 from AS5 (this AS)

output
policy
engine

metric=10
172.16.10.0/24 path=(AS5)

172.16.10.0/24 path=(AS5)
192.213.1.0/24 path=(AS5 AS1)

AS4

Fun with BGP

Routeviews.org collects and archives BGP

announcements
One way to use routeviews is with dig

At the linux prompt

dig txt 4.128.aspath.routeviews.org
Outputs various stuff and

Answer section:
4.128.aspath.routeviews.org 600 IN TXT 5056 1238 174 34
128.4.0.0 16
Syntax = ASPath Prefix prefix length

Now use whois -h whois.arin.net "a ASXX" to learn

about ASs where XX is an AS number. E.g., whois

-h whois.arin.net "a AS34" gives information about
AS34
Try with some other AS

Check out a collection of path

announcements

Open bgp030408p39.Partial

http://www.eecis.udel.edu/~bohacek/Classes/ELEG651Spring2008/bgp030508p39.Partial
An old (2003) partial list of BGP announcements received by several routers

Check which ASs peer with UD (ASN 34)

Why different Intra- and Inter-AS routing ?

Policy:
Inter-AS: admin wants control over how its traffic

routed, who routes through its net.

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