Questions to discuss for the Arpanet Metrics paper. Discuss any one of the issues listed below. Please post your response at http://hydralisk.cs.washington.edu.

1. The first iteration of the system used a Bellman Ford distributed algorithm before finally settling on a link state algorithm. Discuss under what setings you would use Bellman Ford and under what settings under which you would use the alternative link state algorithm.
2. Discuss what are the desirable properties of a routing system.
3. Why are routing oscillations considered harmful?
4. How would you generalize link state routing to work with multiple metrics such as latency and link load? What are the issues in generalizing link state routing with multiple metrics?
5. Speculate regarding the relevance of this paper and the work for today's networks.