CRI: Scalable Embedded Wireless Sensor Network Evaluation Facility

University Of California-Berkeley, Berkeley CA

Investigators

Abstract

Abstract Proposal: CNS 0454432 PI: David E. Culler Institution: University of California-Berkeley Program: NSF 04-588 CISE Computing Research Infrastructure Title: CRI:Scalable Embedded Wireless Sensor Network Evaluation Facility Investigators at UC Berkeley will build an inter-related family of flexible, large-scale evaluation facilities for wireless embedded sensor networks that will enable empirical evaluation of critical protocols at a scale and completeness that is not possible today. This will allow rigorous testing of protocols and distributed data processing algorithms under real conditions of noise, interference, loss, and variations in connectivity that are extremely difficult to capture accurately under simulation. They will build an indoor, large-scale set of over a thousand heterogeneous wireless nodes with extensive logging capabilities. The testbed will utilize the newest generation of Berkeley motes, redesigned for these testbeds, as well as newly developed, embedded Linux-class nodes. This will be complemented with a dedicated server side infrastructure to schedule experiments, operate nodes, store and analyze the experimental data, and emulate enterprise capabilities that interact with sensor networks. Traces of sensor data streams will also be stored in the infrastructure and replayed into the wireless network to allow repeatable experimentation with realistic in-network data processing algorithms. In addition to the sheer number of nodes, the facility will stress the heterogeneity likely to be encountered in practical deployments by including higher-powered nodes, higher bandwidth networks, and mobile motes. The infrastructure will enable research in Wireless Sensor Networks which represent a very new and important class of networked systems. Issues to be addressed include routing, information dissemination techniques, energy efficient communication protocols and reliable communication with intermittent connectivity, the use of vision and high bandwidth sensors, mobile and high power nodes in heterogeneous WSN, and networked information theory. Finally, approaches to security in WSN such as analysis of key pre-distribution protocols, secure routing, secure aggregation and privacy research will be highlighted on the testbed. In addition to supporting about 25 researchers at Berkeley, the testbed network will become available to the broader research community through Internet login or on-site visits. Broader Impacts of the project include exploring technology for eventual deployment of a new information infrastructure to a large number of critical physical infrastructures, including electric power, heating and cooling of buildings, water, oil and gas pipelines, refineries, etc. This project will examine high confidence services: reliable, fault tolerant and resistant to attack. These same capabilities represent a potential breakthrough for many areas of science, including environmental monitoring, civil engineering, eco-physiology and others.