Critique on InterDomain Internet Routing Lecture 4 by H. Balakrishnan

This paper takes a look at how routing is done between different ASs in the Internet. It also present how they ideally exchange route information within the AS and with other ASs, and how they actually do in real world which leads to real world issues. The first part discussed on relationship between AS and classify their relationship base on their behavior in exchange of routing information. It then proceed to the discussion of BGP and its salient features, a widely used wide-area routing protocol today, which is a simple protocol and yet complex to operate. This routing protocol tries to resolves the following issues such as scalability, policy and cooperation under competitive circumstances. The material presents some interesting open issues in policy, failover, scalability, and configuration.

The material is light reading, enjoyable and a good introductory material for research in wide-area routing.

I think the classifications of inter-relationship between AS is insufficient to fully handle all existing relationship between all AS. One of the premise which I don't agree with is the assumption that ASs don't trust each other and tries to limit each others reachability and connectivity optimization. Example, there may exist an ASs having a mutual trust and allows its network to be utilize to transit data of other network when its network is not congested in order to improve the overall end to end performance of the Internet. Their co-existence benefit them both and to the end-user. The trust may be gain by some means and cooperation might occur when there is no benefit for both AS to compete.

Although the assumptions of this paper is statistically correct in our time. It may be changed as we progressed.


References:

[1] Balakrishnan, Hari and Nick Feamster. "Interdomain Internet Routing", 2005.

[2] Peterson, Lary and Bruce Davie. Computer Networks: A System Approach. 3rd ed. San Francisco: Morgan Kaufman, 2003.