OVERVIEW

The Systems Area in CENS focuses on the design and deployment of long-lived self-configuring Embedded Networked Sensing (ENS) systems. With research spanning the spectrum from a study of fundamental properties and algorithms to implementation of practical systems and testbeds, the goal is to develop a comprehensive understanding of how to systematically design, run, and manage ENS systems. These systems are characterized by energy constraints, irregular configurations, time-varying topology, large scale and changing applications, all of which render simplistic hardwired configuration and functions non-scalable. We will use the terms ENS and sensor networks interchangeably.

The overall research effort is divided into three sub-areas: Network Autonomy, Programming and Storage, and Test-beds and tools.

The second thrust, Programming and Storage, focuses on the important problem of how can one raise the abstraction of programming and data storage, and ENS usage from the level of individual nodes to that of a collective or an aggregate. The vision of ENS systems has always been one where a collection of nodes is treated as a single entity, albeit a spatially distributed one, that monitors the phenomena of interest. The reality however is that these systems are still programmed at the level of individual nodes. Projects under this thrust focus on distributed computation and storage frameworks, and their use for tasks such as distributed aggregation. In the past year, we have had significant accomplishments in this area. Several of our more mature efforts have resulted in concrete deliverables. Our SensorWare project resulted in an important publication and, as well, released code to the public domain. Our distributed storage efforts also achieved some important milestones, with the DIM system being integrated with TinyDB and a variant of the DIMENSIONS system being implemented on the motes as a proof-of-concept for structural applications. Equally exciting were the newer developments in this area which are currently in the prototyping stage: the SNACK toolkit which is built on top of nesC and TinyOS, and the MOAP over-the-air programming project have both reported successful prototype developments and will see significant advances in the coming year.

OBJECTIVES

The objective of research in the systems area is the design and deployment of long-lived self-configuring Embedded Networked Sensing (ENS) systems which are characterized by energy constraints, irregular configurations, time-varying topology, large scale, and changing applications. The research spans the spectrum from study of fundamental properties and algorithms to implementation of practical systems and testbeds, and seeks to develop a comprehensive understanding of how to systematically design, run, and manage ENS systems.