TABLE OF CONTENTS Kenneth P. Birman, Saikat Guha and Rohan Murty
Department of Computer Science
Cornell University, Ithaca, NY 14853
{ken,sg266,rnm5} @ cs.cornell.edu
| Abstract | Sensor networks will often need to organize themselves automatically and adapt to changing environmental conditions, failures, intermittent connectivity, and in response to power considerations. We review a series of technologies that we find interesting both because they solve fundamental problems seen in these settings, and also because they appear to be instances of a broader class of solutions responsive to these objectives. |
| Keywords: | Embedded computing, sensors, distributed monitoring, routing, network position information. |
The emergence of a new generation of technologies for pervasive computing and networking is challenging basic assumptions about how networked applications should behave. Wired systems typically ignore power and location considerations and operate "in the dark" with respect to overall system configuration, current operating modes or detected environmental properties, and positions of devices both in absolute and logical terms. These kinds of assumptions represent serious constraints and lead to sub-optimal solutions in embedded or pervasive computing systems.
At Cornell, we and other researchers are working to develop platform technologies responsive to these and related considerations. This paper reports on three representative examples, which we offer with two goals in mind. First, each of these technologies reflects a mixture of properties and algorithmic features matching the special requirements seen in pervasive computing settings. Second, we believe that the underlying methodologies reflected in the three technologies are interesting in themselves, because they point to broader opportunities...
