Building Routing Table
There are 2 methods of building routing tables
.
1. Manually
2. Software
There are then further 2 methods of computing these.
Static Routing
It is not flexible.
It is done at boot time
It has small network overhead.
Dynamic Routing
It can work around network failures automatically.
It allows automatic updates by the programmers.
Computing Shortest Path in a Graph
We use Djikstra's [^1] algorithm to compute the shortest path.
Distributed Route Computation
If a node fails then the switch
updates its routing table information to avoid the failed hardware.
Distance Vector Routing
The switches
periodically broadcast topology information throughout the network.
Then other switches
update their routing tables according to the information received.
Vector - Distance Algorithm
Link-State Routing
The network topology is separated from route computations.
switches
send link-state
information about local connections and then they build their routing tables
Comparison
Distance Vector Routing
- Simple to implement
switch
updates its own routing table first- It is used in
RIP
.
Link State Algorithm
- Complex to implement
switch
performs independent computations.- it is used in
OSPF
.
Example of WAN Technologies
ARPANET
Began in 1960s, it was funded by Advanced Research Project Agency, which is an organization of US defense department.
X.25
This was standard for connection-oriented networking.
It began from IFU which was originally CCITT.
Frame Relay
It is used for Telco service for delivering blocks of data.
It is connection based service and requests Telco to establish the circuit between 2 end points.
Its speeds are following:
- 56 Kbps
- 1.5 Mbps
- 100 Mbps
SMDS
Switched Multi megabit Data Service
is also a Telco service.
This one is connection less and speed ranges from 1.5 Mbps to 1000 Mbps.
ATM
Asynchronous transfer mode
was designed for lower jitter
, high capacity, audio, video and data.
It uses fixed size 48 octets data and 5 octets header.
-
Read more about Djikstra's Algorithm. ↩