Welcome!

Video/Movie On-Demand is an interesting research area that offers a great challenge to work. Yes, designing high fidelity multimedia servers/systems is by no means a trivial issue. Especially, when designing systems that can render services on a network-based infrastructure, the issues are all the more challenging.

I am interested specifically in designing "pull-based" video server systems - a concept that embraces data divide-and-conquer policy for video retrieval. I have conducted some theoretical studies on the design and performance analysis of strategies that help to retrieve long-duration movies from a server system on a network. Yes, I am talking about retrieving long-duration movies (a full-fledged ~110mins English movie!) via a network system. So, at present, our group has an excellent theory that aids a "fast transfer" (your waiting time is minimized to the best possible extent) by a clever design of a scheduler. Further, work pressure on servers are balanced, which in turn aids to maximize the number of clients that can be served by the server system. This is certainly an attractive scheme for a service provider.

Currently, 1 PhD, 1 Meng, and 3 Final Year Project students (FYP 2000-2001) are actively working in this area. We look at several pending issues to be addressed from theoretical perspective (design of algorithms, strategies, etc) and FYPs thrive to implement and realize a "working system". A full-fledged implementation is completed and we are not opening the "source" for some time until "certain formalities" are completed.

If you are interested to join our V/MoD group, you are most welcome. Let us talk to see whether we can bridge our interests, first. Thanks for visiting this page!

Some recent documents (available on request):

1. Access Time Minimization for Distributed Multimedia Applications

Presents strategies for long-duration movie retrieval on delay sensitive distributed networks. Employs retrieval via multiple servers to a single client location.

2. Efficient Movie Retrieval Strategies for Movie-On-Demand Multimedia Services on Distributed Networks

Extension to the above problem. Retrieves a movie via multiple servers using multiple channel retrieval strategy to  multiple clients; presents analysis on buffer occupancy at the client site and derives optimal size distribution from several servers.