7.

Network Layer

The Network Layer is responsible for getting packets from source to

destination, possibly over multiple links
z Primary functions include:
Routing
Congestion Control
Internetwork data exchange
Provision of quality of service (QoS)
z We will concentrate on the issue of routing at the moment.
z The other issues of the network layer will be addressed in context of the Internet

The Transport Layer is responsible for end-to-end data transport

z Primary functions include:
Provision of connection oriented or connectionless service
Disassembling and reassembling data
Setup and release of connections across the network
z Some transport layer functions will be addressed in the Internet section
EE210 Data Communications and Computer Networks
1

Recall the OSI Model

EE210 Data Communications and Computer Networks

2

Network Layer: Routing Algorithms

routing algorithm = used by a router to decide on which output link each incoming
Packet will be transmitted
datagram packet-switching: this decision must be made for every Packet
virtual circuit packet-switching: routing decisions made only at v.c. set-up
desirable properties of a routing algorithm:
correctness, simplicity, efficiency obviously
robustness since usually the entire network cant be re-booted !!!
stability routing algorithm reaches equilibrium in a reasonable time
fairness, optimality
often in conflict, e.g.
link capacities
fully utilised by
A-A, B-B, C-C
traffic respectively

optimality with respect to what? What are we trying to optimise?!

average Packet delay? total Packet throughput?
these goals are often in conflict: operating any network near capacity
implies long queueing delays in node buffers
compromise minimise number of relays (or hops) a Packet needs
EE210 Data Communications and Computer Networks
3

Network Layer: Routing Algorithms (cont.)

least-cost routing:
a value is assigned to each link in the network: this is the cost of using this link
the cost of a route is the combination of the values of its links
the best route is the one with the lowest cost know how to relay incoming Packets

cost assigned to a link could be:

1 for each link finds route with the fewest hops
(financial) cost of using the link finds cheapest route
packet delay on the link finds minimum-delay route
packet transmission time on the link finds maximum-bandwidth route
or some combination of these, or other factors !

routing algorithms can be divided into 2 classes:

nonadaptive, or static routing decisions are pre-determined and not based on
measurements (or estimates) of the current network topology and traffic load
adaptive routing decisions may be changed when network topology and/or
traffic load changes
extreme case: select a new route for each Packet
may get information just from neighbouring routers, or from all routers
may re-determine routes periodically, or when topology changes, or when
traffic load changes more than a threshold percentage, or
EE210 Data Communications and Computer Networks
4

Network Layer: Routing Algorithms (cont.)

once the cost of each link is known, the routers can run a routing algorithm to
determine the best routes for each possible sender-receiver transmission
in practice: routing algorithm should be adaptive and de-centralised
The most common routing algorithms are distance-vector and link-state
distance-vector: each router exchanges information about the entire network
with neighbouring routers at regular intervals
neighbouring routers = connected by a direct link (e.g. a LAN)
regular intervals: e.g. every 30 seconds
link-state: each router exchanges information about its neighbourhood with
all routers in the network when there is a change
neighbourhood of a router = set of neighbour routers for this router
each routers neighbourhood information is flooded through the network
change: e.g. if a neighbouring router does not reply to a status message
as link-state converges faster, in practice is more widely used
converges = determines optimal routes, given a particular network topology
EE210 Data Communications and Computer Networks
5

Network Layer: Distance-Vector routing

cost = 1 for every link finds minimum-hop routes

clouds represent LANs; number in cloud represents network ID

A, B, C, D, E, F are routers (or gateways)
EE210 Data Communications and Computer Networks
6

Network Layer: Distance-Vector routing (cont.)

each router sends its information about the entire network only to its neighbours

how do non-neighbouring routers learn about each other and share information?
a router sends its information to its neighbours
each neighbour router adds this information to its own, and sends the updated
information to its neighbours; the first router learns about its neighbours neighbours
EE210 Data Communications and Computer Networks
7

Network Layer: Distance-Vector routing (cont.)

each router stores information about the network in its routing table
Network ID = final destination of Packet
Cost = number of hops from this router to
final destination
Next Hop = neighbouring router to which
Packet should be sent
initially, all a router
knows is the network
IDs of the networks
to which it is directly
connected

initial routing
table exchanges
(no multi-hop
routes yet)
EE210 Data Communications and Computer Networks
8

Network Layer: Distance-Vector routing (cont.)

how is a routers routing table updated when new information is received?
keep the entry
with lowest cost

as B is 1 hop
from A

routing table update algorithm (distributed Bellman-Ford algorithm):

add 1 to cost of each incoming route (since each neighbour is 1 hop away)
if a new destination is learned, add its information to the routing table
if new information received on an existing destination:
if Next Hop field is the same, replace existing entry with the new
information even if the cost is greater (new information invalidates old)
if Next Hop field is not the same, only replace existing entry with the
new information if the cost is lower
EE210 Data Communications and Computer Networks
9

Network Layer: Distance-Vector routing (cont.)

example of routing table update algorithm (unrelated to earlier Example network):

EE210 Data Communications and Computer Networks

10

Network Layer: Distance-Vector routing (cont.)

final (converged) routing tables for earlier Example network:

Note: choice between

equal-cost routes
depends on exact
sequence of updates
EE210 Data Communications and Computer Networks
11

Network Layer: Distance-Vector routing in practice

original ARPANET (forerunner of the Internet) used distance-vector routing
subsequently used in the Internet as RIP (Routing Information Protocol)
a variation of distance-vector routing is used in BGP (Border Gateway
Protocol), which finds routes from one autonomous system (AS) to another AS
AS = a part of the Internet (e.g. a network) managed by one entity
link cost can be something other than 1 for each link
e.g. packet delay, number of packets queued,
problem with distance-vector routing: count-to-infinity problem
this refers to the slow convergence of distance-vector routing algorithms
under some conditions
basic flaw slow reaction to link/router failure because information only
comes from neighbouring routers and it may be out-of-date (e.g. it may not
properly reflect the impact of the failure on route costs)
many ad-hoc solutions have been tried (e.g. split horizon), but either they
also fail to solve the count-to-infinity problem, or they are hard to implement
this slow convergence was one of the main reasons why other types of
routing algorithm were explored, leading to link-state routing
EE210 Data Communications and Computer Networks
12

Network Layer: Link-State routing

each router sends information about its neighbourhood to every other router:

EE210 Data Communications and Computer Networks

13

Network Layer: Link-State routing (cont.)

link cost is usually a weighted sum of various factors
e.g. traffic level, security level, packet delay,
link cost is from a router to the network connecting it to another router
when a packet is in a LAN (which is typically a broadcast network), every
node including the router can receive it no cost assigned when going
from a network to a router
Note: costs shown
are examples only

EE210 Data Communications and Computer Networks

14

Network Layer: Link-State routing (cont.)

routers share information by advertising, which means sending link-state packets
Advertiser: ID of sending router
Network: ID of destination network
Cost: link cost to neighbour
Neighbour: ID of neighbour router

a router gets its information about its neighbourhood by sending short ECHO
packets to its neighbours and monitoring the responses

the Figure shows

how router As
link-state packet
is flooded to all
other routers

EE210 Data Communications and Computer Networks

15

Network Layer: Link-State routing (cont.)

every router builds a link-state packet and floods it through the network, so when
all such packets have been received at a router, it can build its link-state database:

Assuming that every router

receives the same set of
link-state packets (as if the
routers were synchronised),
every router builds the same
link-state database.
Using this database, each
router can then calculate
its routing table.

EE210 Data Communications and Computer Networks

16

Network Layer: Link-State routing (cont.)

to calculate its routing table, a router uses Dijkstras Shortest-Path algorithm
first, identify all link costs in the network: either from the link-state database,
or using the fact that the cost of any link from a network to a router is 0

EE210 Data Communications and Computer Networks

17

Network Layer: Link-State routing Dijkstras algorithm

this algorithm builds a shortest-path spanning tree for the router: such a tree has a
route to all possible destinations, and no loops
the router running the algorithm is the root of its shortest-path spanning tree
even if all routers link-state databases are identical, the trees determined by
the routers are different (since the root of each tree is different)
a node is either a network or a router; nodes are connected by arcs
the algorithm keeps track of 2 sets of nodes and arcs Temporary and Permanent
initially, the Temporary set contains all neighbour nodes of the router itself,
and the arcs connecting them to the router; only the router is initially Permanent
when all nodes and arcs are in the Permanent set, the algorithm has terminated
identify the Temporary node whose arc has the lowest cumulative cost from the root:
this node and arc are moved into the Permanent set;
any nodes which are connected to the new Permanent node and are not already in the
Temporary set, along with the connecting arcs, are made Temporary. Also, if any node
already in the Temporary set has a lower cumulative cost from the root by using a route
passing through the new Permanent node, then this new route replaces the existing one;
repeat until all nodes and arcs are Permanent.
EE210 Data Communications and Computer Networks
18

Network Layer: Link-State routing Dijkstras algorithm (cont.)

as an Example, lets follow the steps of the algorithm run by router A

1.

2.

3.
Note: arcs are
marked with their
cumulative cost
from the root (not
individual costs)
EE210 Data Communications and Computer Networks
19

Network Layer: Link-State routing Dijkstras algorithm (cont.)

4.
Note: arcs are
marked with their
cumulative cost
from the root (not
individual costs)

5.

Note: equal cumulative costs

choose one arbitrarily in step 6
EE210 Data Communications and Computer Networks
20

Network Layer: Link-State routing Dijkstras algorithm (cont.)

6.

Note: arcs are

marked with their
cumulative cost
from the root (not
individual costs)

7.

EE210 Data Communications and Computer Networks

21

Network Layer: Link-State routing Dijkstras algorithm (cont.)

8.

Note: arcs are

marked with their
cumulative cost
from the root (not
individual costs)

9.

EE210 Data Communications and Computer Networks

22

Network Layer: Link-State routing Dijkstras algorithm (cont.)

10.

Note: arcs are

marked with their
cumulative cost
from the root (not
individual costs)

11.

if the new arc to network 66 from

router D had a lower cumulative cost
than the one from router C, then
the new link would replace the old one
EE210 Data Communications and Computer Networks
23

Network Layer: Link-State routing Dijkstras algorithm (cont.)

12.

Note: arcs are

marked with their
cumulative cost
from the root (not
individual costs)

13.
all nodes and arcs are
Permanent STOP:
this routers shortest-path
spanning tree has been found
EE210 Data Communications and Computer Networks
24

Network Layer: Link-State routing routing table

once a router has found its shortest-path spanning tree, it can build its routing table
to complete the Example, here is router As link-state routing table:
Note: each routers routing table
will (in general) be different
Networks 14, 23, and 78 dont have
a Next router entry because they
are directly connected to this router

in large networks, the memory required to store the link-state database and the
computation time to calculate the link-state routing table can be significant
in practice, since the link-state packet receptions are not synchronised, routers may
be using different link-state databases to build their routing tables: how inaccurate
the results are depends on how different the routers views of the network are
EE210 Data Communications and Computer Networks
25

Network Layer: Link-State routing in practice

link-state routing algorithms have several desirable properties, e.g. rapid convergence; small
amount of traffic generated; rapid response to topology changes

examples from the Internet are the Open Shortest Path First (OSPF) and Intermediate
System to Intermediate System (IS-IS) routing protocols

link costs can be configured in OSPF. Possible link costs include:

1 for each link
reliability: assigned by administrator, indicates how often the link fails
packet delay
link bandwidth
financial cost of the link

OSPF requires a lot of memory: each router holds its routing table & link-state database
Dijkstras algorithm computations are processor-intensive
legacy routers may be unable to relay packets when these calculations are taking
place which could be every time a link-state packet is received

OSPF can consume a lot of bandwidth if the network topology changes often
link-state packets sent to all routers using reliable flooding
need sequence number and time-to-live (TTL) field in each packet
EE210 Data Communications and Computer Networks
26