Probabilistic addressing: Stable addresses in unstable wireless networks

Unreliable connectivity and rapidly changing link qualities make it challenging to establish stable addressing in wireless networks. This is especially difficult in communication scenarios where nodes determine their own addresses based on the underlying connectivity in the network. In this paper, we present Probabilistic ADdressing (PAD), a virtual coordinate based addressing mechanism that efficiently deals with dynamic communication links in wireless networks. It assigns probabilistic addresses to nodes without needing to pessimistically estimate links over longer periods of time. The routing metric predicts the current location of a node in its address distribution. Our prototype implementation over real testbeds compares PAD with Beacon Vector Routing (BVR), a prominent routing approach in sensornets, and S4, a cluster based extension of BVR and state-of-the-art in point-to-point routing in sensornets. Our results from three widely used testbeds indicate that PAD achieves 3-7 times more stable addressing than BVR, and minimizes the number of transmissions in the network by 26% when compared with S4 under challenging networking conditions.