Overview
In 1943, Thomas J. Watson, Chairman of IBM declared: “I
think there is a world market for maybe five computers.” A few billion
computers later, it may be tempting to declare that "there is a world market
for maybe five computers per capita per year." But this is not how computer
science visionaries see the world in the twenty first century. Computer
scientists envision a world in which computing is pervasive and seamless.
Large-scale distributed, systems composed of countless sensors and actuators
embedded in the physical world, will eventually infuse the entire world, but at
a physical level instead of the virtual level to which we have been accustomed
in our short experience with the Internet and the Web. These sensors and
actuators may not do much individually. But, when networked, they have the
potential of performing quite elaborate functions. Thus, what these embedded
devices will be capable of doing (and consequently how they will transform the
way we live) is limited only by our ability to harness their collective power
through:
-
Protocols and services that recognize and tolerate the
constraints under which these devices will be embedded in the physical world
(including ad-hoc deployment, power constraints, transient failures and noisy
readings, etc.)
-
Distributed algorithms for the in-network processing and
management of the huge amounts of sensory data collected by these devices
This seminar aims to introduce its participants to the
various "stakeholder areas" in sensor networks. As such, this seminar will be
team-taught by a group of faculty members (Bestavros,
Byers,
Kollios,
Matta, and
Teng) in the sense that each faculty
member will oversee a particular dimension of sensor networks.
This seminar will run throughout the 2003/04 academic year.
It will be offered as a 2-credit course in each of the Fall and Spring
semesters. For the Fall semester, the class will meet once weekly on Tuesdays 2:00pm-3:30pm in
SAR 218.
Goals and Topics
The following is a preliminary list of
"dimensions" or focus areas (many of which are intertwined) that this seminar
may examine--dependent on interest from its participants. The reading list
(organized under the dimensions below) is available on a
separate page.
-
Introduction
-
Visions and dreams!
-
Applications
-
Taxonomies
-
Design and Implementation Issues
-
Evaluation Issues
-
Topological Issues
-
Data Manipulation Issues
-
Storage, Indexing and Querying
-
Abstract Data Structures
-
Dissemination and Diffusion
-
In-Network Stream Manipulation
-
Coordination Issues
-
Cross-cutting Issues
Eligibility and Prerequisites
This seminar will run throughout the 2003/04 academic year. It will be offered
as a 2-credit course in each of the Fall and Spring semesters.
In the Fall semester, this course will be open to all graduate (PhD, MA, and BA/MA) students in the
Computer Science Department. All other students must obtain the explicit
approval of one of the faculty members co-teaching this seminar (Bestavros,
Byers,
Kollios,
Matta, and
Teng). This course is not available
for auditing, except if approved by a faculty member co-teaching the class.
The following courses (or equivalents) are required
-
CS-350: A solid understanding of elementary aspects of computing
systems
- CS-455: A solid understanding of basic networking concepts and
architectures
In addition to the above courses, the following
courses are highly recommended in the sense that
they will leverage a student's ability to pursue the required course work.
-
CS-410:
Basic understanding of system
architecture and programming
- CS-556: Familiarity with
advanced networking tools and techniques
- CS-562: Manipulation, indexing, clustering and summarization
of large data sets
Note: In the Spring semester, this course will be open only to
students who secured a A grade (A or A-) in the Fall offering of the course. |