PhD Thesis Proposal Defence "Overlay Topology Inference and Tree Construction" by Mr. Xing Jin Abstract: In the recent years, overlay networks have been increasingly used to deploy network services. In order to build an efficient overlay network, the knowledge of underlay is important. In this proposal, we study the following two problems: (1) How to efficiently infer the underlay topology among hosts by means of end-to-end measurement tools such as traceroute? We propose a Max-Delta heuristic to infer a highly accurate topology among hosts with a few traceroutes. We further study how to deal with the noise of anonymous routers in measurements. (2) Given an underlay topology, how to build a high-bandwidth overlay tree among hosts? We formulate the problem as building a Maximum Bandwidth Multicast Tree or a Minimum Stress Multicast Tree, depending on whether the link bandwidth is available or not. We prove that both of them are NP-hard and not approximable within a factor of 2/3+e, for any e>0, unless P=NP. We then present and analyze approximation algorithms to address them. We have evaluated the proposed schemes on Internet-like and real Internet topologies. Our results show that almost full measurement is needed to fully discover the underlay topology. However, substantial reduction in measurements can be achieved if a little accuracy, say 5%, can be compromised. And our merging algorithms in the presence of anonymous routers can efficiently infer an underlay topology with good accuracy. Furthermore, as compared to traditional tree-based protocols such as Narada and Overcast, our approximation algorithms achieve high tree bandwidth and low link stress with low penalty in end-to-end delay. Our study show that indeed the knowledge of the underlay is important for constructing efficient overlay trees. A limitation of above approaches is that the schemes rely on a central server to collect traceroute results and to construct the overlay tree. In the future, we will study distributed topology inference and distributed tree construction techniques. Date: Tuesday, 6 February 2007 Time: 10:30a.m.-12:30p.m. Venue: Room 3501 lifts 25-26 Committee Members: Dr. Gary Chan (Supervisor) Dr. Jogesh Muppala (Chairperson) Dr. Brahim Bensaou Prof. Lionel Ni **** ALL are Welcome ****