Utility Maximization for
P2P Applications and Its Application in Optimizing Multi-party Conferencing
Jin Li
Microsoft Research
Redmond, WA
Friday, April 16, 2pm
Location: BAB024 (Bahen Centre Basement)
Abstract:
Peer-to-Peer (P2P)
applications are tremendously popular on today's Internet, and applications
like file sharing perform well even with thousands or millions of peers.
However, the design of the majority of P2P systems does not strive to achieve
any systematic optimization of social welfare to all peers under a resource
sharing constraint. This may well be the next step in improving the
performance of P2P systems.
In this talk, we study the problem of utility maximization in P2P
applications, in which aggregate application- specific utilities are maximized
by running distributed algorithms on P2P nodes. Using recent discoveries in
optimal tree structure for P2P content delivery, we develop a new formulation
for multicast utility maximization problem. This formulation is unique in the
sense that it not only eliminates some mathematical difficulties as compared
to previous formulations, but also leads to practical solutions. We develop
Primal and Primal-dual distributed algorithms to maximize the aggregate
utility. We prove that these algorithms converge to the optimal solution of
the utility maximization problem exponentially fast. Furthermore, they can be
implemented by utilizing only the end-to-end delay measurement between P2P
nodes. As such, it can be readily deployed on today's Internet. To support
this claim, we have implemented the Primal-dual algorithm to design a
peer-assisted multi-party conferencing system and evaluated its performance
through actual experiments on a small scale testbed, as well as on the
Internet.
Bio:
Dr. Jin Li
is currently a Principal Researcher managing the Communication System team at
Microsoft Research, (Redmond, WA). He received his Ph.D. in electrical
engineering from Tsinghua University (Beijing, China) in 1994. From 1994 to
1996, he served as a Research Associate at the University of Southern California
(USC). From 1996 to 1999, he was a Member of the Technical Staff at the Sharp
Laboratories of America (SLA), (Camas, WA), and represented the interests of SLA
in the JPEG2000 and MPEG4 standardization efforts. He joined Microsoft Research,
first as a Project Leader at Microsoft Research Asia (Beijing, China) from 1999
to 2000, and then moved back to Redmond in 2001. From 2000, Dr. Li has also
served as an Adjunct Professor in the Electrical Engineering Department,
Tsinghua University (Beijing, China).
Dr. Li has 100+ referred conference and journal papers in a diversified research
field, with interests cover audio/image/video compression, virtual environment
and graphic compression, audio/video streaming, and VoIP and video conferencing,
and P2P networking. His research group (Communication System) supports Microsoft
real-time communication (RTC) group, which develops the audio/video engine for
Microsoft Unified Communication systems. His invention has been integrated into
many Microsoft products, such as Microsoft Office Communicator, Microsoft Live
Meeting, Live Messenger, Live Mesh, Windows 7 (Teredo), etc.. He holds 20+
issued US patents, with more than three dozens pending. He was the general chair
for 17th International Packet Video workshop 2009 (PV 2009), and was on the
organization committee/TPC/associate editors on many conferences and journals.
He was the recipient of the 1994 Ph.D. thesis award from Tsinghua University,
the 1998 Young Investigator Award from SPIE Visual Communication and Image
Processing, and the Best Paper Award from 2009 IEEE International Conference of
Multimedia.
Host of Talk:
Baochun Li (bli@eecg.toronto.edu)