A Self-Adaptive Middleware for Efficient Routing in Distributed Sensor Networks

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.