[EdCert previous] [EdCert next] [EdCert top]

Routing

Routing is what makes true networking possible (and also what is "breaking" the internet today). Routers are rack mounted thingies from companies such as Cisco or just computers with multiple ethernet interfaces which perform the task of routing TCP/IP packets to their destination.

Routing can be static or dynamic. Although routing tables can be built and maintained statically (at least a "default route" is often configured statically on a Unix host), an ever evolving network can easily make the administrator's job hell. Dynamic routing solves this problem and allows network hosts to automatically determine the best routes to remote destinations.

Though there are number of different dynamic routing protocols in use today, RIP is a protocol still used often. The Unix RIP routing daemon is called routed. (O'Reilly's ``TCP/IP Network Administration'' by Craig Hunt offers more information about RIP and other routing protocols.)

For RIP to work, ifconfig is used on a Unix host (as part of configuring the ethernet interface at bootup) to set up an IP broadcast address. RIP broadcasts (made every 90 seconds by routers using this address) go to every host on the attached subnets. Each host on these subnets responds to this broadcast by telling the originator its name/address and how many router hops away it is (this updated information becomes part of the originator's routing table). When a subsequent request is made by network client software, updated routing tables are looked up and a network connection is made appropriately.




[EdCert previous] [EdCert next] [EdCert top]