UC Davis

Routing in sensor networks with unpredictable network connections and node failures is challenging due to the lack of knowledge about network dynamics in decision making. We present a self-adaptive middleware for efficient routing in distributed sensor network. At the heart of the proposed approach is the design of a policy-driven language to control the relocation of software components between sensor nodes. Accordingly, various changes in sensor networks can be dynamically adapted. Based on such a middleware, we provide a few approaches in building a practical routing protocol. For instance, nodes can adaptively switch their routing strategies according to the network stability or migrate some tasks to other idle nodes to prevent from node failures.