|
As Simson Garfinkel wrote in Technology Review, peer-to-peer is much more than those
file-sharing services violating the copyright of big record labels, it is really
how the Internet was originally designed to work.
The peer-to-peer theory is that all of the computers on the network would be
first-class citizens, each capable of sharing resources or exchanging
information with one another. For example, a student at Vanderbilt might start
typing on a computer in Nashville and use it to log into a computer at Berkeley.
Meanwhile, another student at Berkeley might use that same computer to log into
the first system at Vanderbilt. Both computers would be simultaneously using
and offering services to the network. The connections between them would be links
between equals-that is, peer-to-peer.
Although, as it turns out, most of the Internet become server-client system.
In recently years, peer-to-peer has been proved sucessful through several
largescale commercial deployments. Its effectiveness and scalabilities made it its
fame in multimedia streaming applications, such as video on demand, and live
event broadcasting.To deliver satisfied user experience, any peer-to-peer streaming solution must provide
(1) enough bandwidth capacity to support high-quality streaming and
(2) relative stability to minimize the service disruption.
It is easy to satisfy one of them, hard to reconcile both.
Towards this end, we proposed PRW (peer resilient weight), a
metric to evaluate a peer's healthiness regarding these two
aspects. The theoretical background of PRW originates from the
optimization framework based on the generalized flow theory.
It generalizes the classical network flow problem by specifying a gain factor for each link in the network.
In our case, we define this gain factor as a peer's
resilience factor, which means, its chance of survival within a certain
time horizon. As such, the optimization problem becomes
to maximize the aggregated generalized flow received by all
peers.
To show PRW's applicability to a wide variety of P2P streaming
solutions, we propose several parent selection algorithms guided
by PRW under the single-tree, multi-tree, and mesh structures.
Using MSN and PPLive traces, we find our algorithms to be
able to maintain consistently low service disruption and low peer
rejection, compared to several existing heuristics.
See More Detail... [HERE]
This work was supported by the
National Science Foundation (NSF award number 0643488) CAREER: Achieving Self-Tunability of Peer-to-Peer Streaming Service through User-Level QoS Inference. |
|
-
Minimizing Service Disruption in Peer-to-Peer Streaming
Proceedings of INFOCOM, 2009.(Submitted)
- Bo Liu, Yanchuan Cao, Yi Cui and Yuan Xue
[tech report]
-
Optimizing P2P Streaming Throughput under Peer Churning
Multimedia Systems Journal(MMSJ), 2008.
- Yi Cui, Yanchuan Cao, Liang Dai and Yuan Xue
[tech report]
-
BitTube: Case Study of a Web-based Peer-Assisted Video-on-Demand System
Proceedings of ISM, 2008.
- Bo Liu, Yi Cui, Bin Chang, Ben Gotow and Yuan Xue
[pdf][ppt]
|
