Routing Principles: Difference between revisions
No edit summary |
No edit summary |
||
Line 6: | Line 6: | ||
In computer networks routing is needed in lots of different situation. For example in a circuit switched network (like the phone network) a connection needs to be established at the beginning and it remains static till the end of the conversation. In a packed switched network (like the internet), each packet (chunk of information) gets routed seperately, so the decision which way to take needs to be performed way more often. |
In computer networks routing is needed in lots of different situation. For example in a circuit switched network (like the phone network) a connection needs to be established at the beginning and it remains static till the end of the conversation. In a packed switched network (like the internet), each packet (chunk of information) gets routed seperately, so the decision which way to take needs to be performed way more often. |
||
Nodes in the network can be added or removed, therefore the routing needs to adjust dynamically to changes of the network topology. '''Routing algorithms''' calculate the ways we can take through a network on demand and save or update them to '''routing tables'''. |
|||
We distinguish between 2 different types of routing: |
|||
* Global information, where all routers know the whole topology of the network and calculate the routing tables with a '''link state algorithm'''. |
|||
* Decentralized information, where the routers only know their directly connected neighbours and calculate their routing tables with '''distance vector algorithms'''. |
|||
== Link State Routing == |
== Link State Routing == |
Revision as of 15:19, 19 July 2007
Routing Principles
Networking
Routing can be compared with the kind of routing we would do when we try to travel to another city by car. We need to figure out which way to take, we need to get a new route when something on our normal route is blocked, etc.
In computer networks routing is needed in lots of different situation. For example in a circuit switched network (like the phone network) a connection needs to be established at the beginning and it remains static till the end of the conversation. In a packed switched network (like the internet), each packet (chunk of information) gets routed seperately, so the decision which way to take needs to be performed way more often.
Nodes in the network can be added or removed, therefore the routing needs to adjust dynamically to changes of the network topology. Routing algorithms calculate the ways we can take through a network on demand and save or update them to routing tables.
We distinguish between 2 different types of routing:
- Global information, where all routers know the whole topology of the network and calculate the routing tables with a link state algorithm.
- Decentralized information, where the routers only know their directly connected neighbours and calculate their routing tables with distance vector algorithms.