SCADDS: Scalable Coordination Architectures for Deeply Distributed Systems


SCADDS is new research project at USC/ISI sponsored by DARPA under the Sensor Information Technology (SenSIT) program. The SCADDS principal investigators are Prof. Deborah Estrin, Prof. Ramesh Govindan and Prof. John Heidemann.

The SCADDS project explores scalable coordination architectures for deeply distributed and dynamic systems ( e.g. wireless sensor networks) . Nodes in these systems will be heterogeneous, having a range of sensing, actuation and communication capabilities. Many deeply distributed systems will require nodes that are small, low-power, mobile, and wireless. In such systems, nodes lose their individuality. For example, a large sensor net application should not be designed to answer questions like: What is the temperature reading at sensor #27 now? Instead, the DATA generated by these devices should be the focus of all communication, independent of which node collected the data originally, and which nodes currently store it. This paradigm shift requires a different coordination architecture, motivated and elaborated upon in greater detail in our Mobicom 99 research challenges paper, "Scalable Coordination in Sensor Networks".

In particular, we focus on directed diffusion techniques, adaptive fidelity algorithms, and the system building blocks such as localization, time-synchronization and self-configuration needed to support these. More details on these and other SCADDS research can be obtained from our presentations and publications.

We are also  developing several testbeds for evaluating our above algorithms, based on the notion of tiered architectures, elaborated here.


Main    Details     People     Presentations and Publications     Testbeds and Software     Internal page

Last modified on: February 22, 2002

Mail any comments, suggestions to Dawn Johnson.