All seminars are held in Room 337 Towne from 11:00am - 12:00pm.
Date:
02/02/04
Name:
Prof. Ramesh
Rao
Institution:
University of
California at San Diego
Title:
Recent Results in Energy Efficient Communications
Date:
02/23/04
Name:
Prof. Steven
Weber
Institution:
Drexel
University
Title:
Rate Adaptive Multimedia Streams: Optimization, Admission Control, and
Distributed Algorithms
Date:
03/01/04
Name:
Prof. Sergio
Servetto
Institution:
Cornell
University
Title:
Network Flow Methods in Information Theory
Abstract and Biography
Date:
03/15/04
Name:
Prof. Sanjay
Shakkottai
Institution:
University of
Texas at Austin
Title:
Grid Sensor Networks: Statics and Dynamics
Date:
03/29/04
Name:
Dr. Dave Wang
Institution:
Wide Area
Network Design Lab
Title:
MPLS Traffic Engineering and Fast ReRoute Protection
Date:
04/05/04
Name:
Prof.
Srinivasan Seshan
Institution:
Carnegie Mellon
University
Title:
To be announced
Abstract and Biography
Date:
04/12/04
Name:
Prof. Rick Blum
Institution:
Lehigh
University
Title:
Optimum MIMO
Signaling for Antenna Selection and Interference
Prof. Ramesh Rao (University of California at San Diego)
Recent results in Energy Efficient Communications
Abstract:
We first present an overview of the recently established California
Institute for Telecommunications and Information Technology with a special
focus on recently launched research initiatives. We then describe some
recent results on efficient battery utilization. We elaborate on ways to
exploit charge recovery as a mechanism to enhance the capacity of a
battery. The charge recovery phenomenon can be observed under bursty or
pulsed current discharge: if a cell is allowed to relax long enough after
delivering a pulse, recovery takes place at the electrode. The bursty
nature of many data traffic sources suggests that data
transmission/reception in wireless communication devices may provide
natural opportunities for charge recovery. Experimental results that
corroborate analytical and simulation findings will also be described.
Biography:
Ramesh Rao is a member of the faculty of the Department of Electrical and
Computer Engineering at the University of California, San Diego, where he
is currently Professor and Director of the San Diego Division of the
California Institute of Telecommunications and Information Technology
[Cal-(IT)²]. He is the former director of the UCSD Center for Wireless
Communications. His research interests include architectures, protocols
and performance analysis of computer and communication networks, and he
has published extensively on these topics. As Director of the San Diego
Division of Cal-(IT)², he leads several interdisciplinary, collaborative
projects.
Prof Steven Weber (Drexel University)
Rate Adaptive Multimedia Streams: Optimization, Admission Control, and Distributed Algorithms
Abstract:
Streaming multimedia is an increasingly important Internet application.
Two salient features of streaming multimedia applications are their
diversity and their adaptivity. Multimedia streams are diverse in that
their duration may range from minutes to hours, and their average rates
may range from tens of kilobits per second to megabits per second. Streams
are adaptive in that media compression allows for a graceful degradation
in media quality in order to achieve a lower average transmission
rate.
This adaptivity is critical for streaming multimedia to function adequately on best-effort networks like the Internet because it allows for streams to dynamically adapt their transmission rates in response to network congestion, or lack thereof. Several questions are natural to ask:
Allocation
When bandwidth becomes scarce, how should we decide which streams should
reduce their subscription level to mitigate the onset of congestion?
Similarly, when more bandwidth becomes available, which streams should
increase their subscription level?
Compression
How should content providers encode media information so as to offer
clients the greatest flexibility to dynamically adjust their subscription
levels?
Implementation
What types of distributed mechanisms/algorithms can be feasibly
implemented to provide streaming clients with the highest possible Quality
of Service?
Performance
Given a definition of Quality of Service, can we assess the type of
performance a client is likely to experience in terms of system parameters
and stream characteristics?
We address these questions in detail and draw some conclusions which may be surprising. In particular, we will discuss:
Optimization
We identify the optimal adaptation policy as one which grants preferential
treatment to small volume streams, where stream volume is the stream
duration times the maximum transmission rate. Moreover, we demonstrate
that at most two subscription levels are needed under the optimal policy.
Admission Control
We discuss why the optimal adaptation policy is not practical for
implementation, but demonstrate that a simple multi-class admission
control policy, whereby streams do not employ dynamic adaptation, achieves
the same Quality of Service as that achieved under the optimal adaptation
policy.
Distributed Algorithms
We propose a simple distributed algorithm where streams react to network
congestion and probe the network for available bandwidth in a volume
dependent manner. We contrast the performance of this algorithm with a
popular volume independent algorithm and show that increased performance
is possible under the former.
Biography:
Steven Weber received his B.S. degree in 1996 from Marquette University in
Milwaukee, WI, and his M.S. and Ph.D. degrees from The University of Texas at
Austin in 1999 and 2003 respectively. Dr. Weber joined the Department of
Electrical and Computer Engineering at Drexel University in 2003 where he is
currently an assistant professor. Dr. Weber's research interests are centered on
computer networks. Specific interests include rate adaptive multimedia streams,
distributed algorithms on networks, supporting real-time services over
best-effort networks, and pricing mechanisms on networks.
Prof Sanjay Shakkottai (University
of Texas at Austin)
Grid Sensor Networks: Statics and Dynamics
Abstract:
In this talk, we consider a network of sensor nodes in a grid topology, and
consider both static properties as well as network
dynamics. The first part of this talk focuses on structural properties such as
connectivity and coverage of unreliable sensor grids, and the implication of
such properties on routing and surveillance. The second part considers
routing dynamics, where search strategies are considered for networks with
little infrastructure support. Asymptotic analysis based on diffusions is used
to show that strategies which utilize memory distributed over space are
exponentially more powerful
than those that do not use memory.
Biography:
Sanjay Shakkottai is an Assistant Professor in the Department of Electrical and
Computer Engineering at The University of Texas at Austin. He received his Ph.D.
from the University of Illinois at Urbana-Champaign in 2002, where he received
the Nokia research
scholarship and the Tellabs fellowship. His research interests include ad-hoc
and sensor networks, scheduling in wireless networks
and Internet congestion control.
Dr. Dave Wang (Wide Area Network Design Laboratory)
MPLS Traffic Engineering and Fast ReRoute Protection
Abstract:
As MPLS technologies are getting mature and are being deployed in many
carrier's network, the applications keep growing. The number one application is
VPN, followed by TE and FRR. In this talk, I will discuss TE and FRR, focusing
mainly on the details of routing, its interaction with layer 3 IP forwarding,
and many interesting and very practical challenges. Routing details involves:
CSPF, RRR, RSVP reservable bandwidth, Affinity/Mask/Admin Group/Color, Dynamic
vs Explicit routing, Strict vs Loose routing, Load balancing, setup and holding
priority, primary/backup/standby LSP, SRLG (Shared Risk Link Group), etc. We
will also go over Inter-area TE, DiffServ TE as well as several flavors of FRR
protection such as path protection or link protection or node protection.
Biography:
Dr. Dave Wang received his BS in mathematics from National Taiwan University in
1969 and his PhD in mathematics from Massachusetts Institute of Technology in
1973. From 1978-86, Dr. Wang worked as a member of the technical staff at
AT&T Bell Laboratories and Bellcore. Since joining WANDL, Inc. (Wide Area
Network Design Laboratory) as a charter member in 1986, Dr. Wang has been a
major influence to the design and development of NPAT (Network Planning and
Analysis Tools). Towards this end, Dr. Wang has considerable expertise in
the design of ATM, Frame Relay, TDM, and IP/MPLS based networks. He has
worked closely with hardware vendors on routing and CAC details, and with
carriers, ISPs, and enterprise customers on countless network designs and RFPs,
advising them in the planning and analysis of WANs of all sizes, and enhancing
the NPAT tool to meet customers' needs.
Prof. Rick Blum
(Lehigh University)
Optimum MIMO
Signaling for Antenna Selection and Interference
Abstract:
Employing transmit and receive antenna arrays to form multiple-input
multiple-output (MIMO) channels in wireless communication systems is a topic
which has attracted a great deal of attention lately. In standard single
link fading channels employing N transmit and N receive antennas, it has been
shown that the capacity grows approximately linearly with N. In cases
without channel state information at the transmitter, this capacity is achieved
by employing a scaled identity signaling covariance matrix. We consider
two slightly different settings, and for both, we demonstrate that the optimum
signaling covariance matrix is not generally a scaled identity matrix. First we
consider a case where a group of interfering users employ single user detection
for flat Rayleigh fading channels with independent fading coefficients for each
path. When the interference is sufficiently strong a scaled identity
signaling covariance matrix is not best. One promising approach for
reducing complexity of MIMO systems is to use a large number of antennas but a
much smaller number of expensive RF processing chains. In the second setting,
when only the best set of antennas are selected at the transmitter and receiver,
again a scaled identity signaling covariance matrix is not generally best.
Finally we comment briefly on cases where channel state information is available
at the transmitter and show similar findings.
Biography:
Rick S. Blum received a B.S.E.E. from the Pennsylvania State University in 1984
and his M.S. and Ph.D in E.E. from the University of Pennsylvania in 1987 and
1991. Since 1991, he has been with the ECE Dept. at Lehigh University
where he is currently a Professor. He has been an associate editor
for the IEEE Transactions on Signal Processing and for IEEE Communications
Letters.