MPhil Thesis Defence


"Efficient Search and Scheduling in P2P-based Video-on-Demand Streaming
Service"

By

Mr. Huicheng Chi


Abstract

With the widespread deployment of broadband access, Video-on-Demand (VoD)
streaming on the Internet has received increasing attention recently. In
VoD streaming service, videos can be delivered to asynchronous users with
low delay and VCR-like operation support (e.g., pause, fast-forward, and
rewind). This thesis studies how to provide VoD streaming service to a
large population of clients using peer-to-peer (P2P) approach. In P2P
systems, cooperative peers self-organize themselves into overlay networks
via unicast tunnels. Each peer not only receives video data as a client,
but also provides streaming to the other peers as a server. This
cooperative sharing effectively reduces the workload on the server end and
the network throughput utilization is greatly improved.

In a P2P-based VoD system, clients request asynchronous streaming parts
and the buffered contents are continuously changing, so the first
important step is to find suitable partners with the expected data. After
that, it is also very important and challenging for the peers to
collaborate with each other for efficient content delivery, because the
buffered contents are limited and the network bandwidth is highly dynamic.
Thus, the partner search and cooperative scheduling are two key components
in the design of P2P-based VoD system. In this thesis, we propose a
generic buffer-assisted search (BAS) scheme to improve partner search
efficiency by reducing the index structure. We also develop a novel
scheduling algorithm based on deadline-aware network coding (DNC) which
leverages the network coding techniques to improve the network throughput
utilization in data exchange. The DNC algorithm avoids the delay drawback
of network coding by considering the users' play deadline in the
scheduling.

Extensive simulation results demonstrate that BAS can provide faster
response time with less control cost than the existing search methods, and
DNC achieves high streaming continuity even in tight network conditions.


Date:				Thursday, 1 June 2006

Time:				10:30a.m.-12:30p.m.

Venue:				Room 5508
				Lifts 25-26

Committee Members:		Dr. Qian Zhang (Supervisor)
				Dr. Jogesh Muppala (Chairperson)
				Dr. Bo Li


**** ALL are Welcome ****