Supporting Delay-Sensitive Applications in Next-Generation Wireless Networks -- A New Theory Combining Large Deviations with Lyapunov Stability

Speaker:	Dr. Xiaojun Lin
		Department of Electrical and Computer Engineering
		Purdue University

Title:		"Supporting Delay-Sensitive Applications in
		 Next-Generation Wireless Networks --
		 A New Theory Combining Large Deviations with
		 Lyapunov Stability"

Date:		Monday, 14 December 2009

Time:		3:00pm - 4:00pm

Venue:		Room 3412 (via lifts 17/18), HKUST

Abstract:

The trend of wireless communication networks is to move towards fully
packet-based wireless networks that can support both data applications and
high-rate delay-sensitive applications.  The key to the success of such
networks is to be able to provide the stringent delay guarantees that
delay-sensitive applications require, and to maximize the capacity of the
system at the same time. Unfortunately, analyzing delay-performance and
providing delay-guarantees in wireless networks have long been a very
difficult problem due to both the scarcity of the wireless resources and
the unpredictable channel-variation.

In this talk, we focus on using large-deviations theory to study the
delay- and queueing- performance of wireless networks operating under
queue-length-based scheduling algorithms. Although large-deviations theory
allows us to focus on the (simpler) asymptotic regime with small
queue-overflow probabilities, it still leads to a complex
multi-dimensional calculus-of-variations problem that is very difficult to
solve. Hence, analyzing the large-deviations delay-performance and
optimality remains a difficult challenge. In our recent work, we develop a
new theory that combines large deviations with Lyapunov stability to
circumvent this difficult. We show that if a control policy minimizes the
drift of a Lypaunov function, then the policy is optimal for maximizing
the large-deviations decay-rate of the overflow probability that
corresponds to large Lyapunov function values.  We will demonstrate how
this result can be applied to both packet-based cellular networks and
multi-hop wireless networks to easily draw conclusions on the
delay-performance of existing wireless control policies, and to design new
delay-optimal control algorithms.


*****************
Biography:

Xiaojun Lin received his B.S. from Zhongshan University, Guangzhou, China,
in 1994, and his M.S. and Ph.D. degrees from Purdue University, West
Lafayette, Indiana, in 2000 and 2005, respectively. He is currently an
Assistant Professor of Electrical and Computer Engineering at Purdue
University.

Dr. Lin's research interests are in the analysis, control and optimization
of large and complex wireless and wireline networks.  He received the IEEE
INFOCOM 2008 best paper award and 2005 best paper of the year award from
Journal of Communications and Networks.  His paper was also one of two
runner-up papers for the best-paper award at IEEE INFOCOM 2005. He
received the NSF CAREER award in 2007. He was the Workshop co-chair for
IEEE GLOBECOM 2007, the Panel co-chair for WICON 2008, and the TPC
co-chair for ACM MobiHoc 2009.