Douglas E. Comer

Computer Science Department
Purdue University
West Lafayette, IN 47907

webmaster: W. David Laverell

Companion Topics

Home

Book Information

Purpose of Site

Getting Started

Students

Faculty

"What's New"

"Coming Attractions"

"C Pointers"

Acknowledgements

Experiment 17.3 - Configure and Run RIP Software on an Emulab

This looked like an easy one, but it turned out not to be so. The first decision is to use FBSD-410.STD. This gives you easy access to routed. You can use the same script as in the basic routing example in which you do all the routing. This gives you three nodes, node0, node1, and node2 with links between node0 and node1 and between node1 and node2.

Emulab normally sets up routing with the rtproto command. Just change Static Manual so that you can be adding routes yourself.

At this point there is something very much different between the Calvin Emulab and the original at Utah. If you try to ping node2 from node0 at Utah, you get what you would expect, nothing! At Calvin, the ping packet tries to go out the gateway and gets filtered. Same result: it doesn't get an answer, but Utah's setup is much better. The Emulab at the University of Kentucky gives similar results to those at Utah.

Now do the following:

• type sysctl -w net.inet.ip.forwarding=1 on all nodes (this enables routing).

• type routed -s -P rdisc_interval=15 on each node.

• give it a little time, and node0 and node2 will be able to ping each other.

• It is fun to kill routed on node1 and see how long it will take before nodes 0 and 2 can no longer talk to each other. That is a function of the parameter 15. Then restart on node1, and see how long it will take before nodes 0 and 2 reestablish contact.

Once you have routed running on all machines you can use netstat to examine the routing table on node0 and fulfill Experiment 17.1 as a bonus.

This is a fun lab once you get it going.