12:45PM, Thursday, December 1st 2005.
Gates 104

Pools, Power, and Interference: Defining a Sensor Network Architecture

Philip Levis
Stanford University

About the talk:
In the beginning, sensor networking knew no boundaries. Unconstrained by protocol layers, system researchers explored a wide range of of architectures, decompositions, and abstractions. The ability to easily define new interfaces and complete network stacks produced a diverse set of protocols and services as well as a meaningful investigation into system requirements. While this flexibility was critical for early progress, today it is a limitation. Existing protocols and subsystems make a wide range of assumptions about the underlying system, and in practice these assumptions tend to be mutually exclusive. Unable to build on existing work, applications start from ground zero and end up being monolithic and vertically integrated designs.

In this talk, I make the case for a sensor network architecture, argue why it must be distinct from the Internet architecture, and propose an architecture that meets the requirements that sensor networks introduce.  Major departures include the relationships between the data link and network layers, transport, storage, link behavior, and traffic patterns. A task of this magnitude is very much a work in progress, and what I discuss is joint work with UC Berkeley.

About the speaker:
Philip Levis is an Assistant Professor of Computer Science and Electrical Engineering at Stanford University. He researches wireless sensor network systems, protocols, applications, and programming methodologies.