SCIMPS: an integrated approach to distributed processing in sensor webs

Mixed signal, micro-electro-mechanical, and wireless communication technologies have accelerated our ability to define and build existing and newly emerging hardware platforms into globally and geographically distributed virtual computing systems. The application domain of these new sensor webs is broad, ranging from bio-medical applications, through remote environmental analysis and sensing, to bio-terrorism. Limited hardware and software, low bandwidth, limited power supply all put numerous strict demands on the design of communication and computation models for the sensor web systems. We discuss a new tightly coupled computation/communication design (SCIMPS), and it's supporting communication protocol KUASAR. KUASAR was developed to support the unique operational requirements of sensor webs, including ad hoc deployment of unknown numbers of processors, and data accessing based on attributes and abstract domains. We have successfully simulated KUASAR on TOSSIM, a simulator developed for Berkeley Motes, and have performed a formal verification of the protocol using SPIN.