Smart Gossip: Infusing Adaptivity into Gossiping Protocols for Sensor Networks

Probabilistic techniques have been used to address many challenges in sensor networks. However, little work exists on developing adaptive versions of probabilistic protocols. In this paper, we consider the class of probabilistic gossiping protocols that are useful for disseminating information. Information dissemination is often required in sensor networks for code updates, TAG-type queries, etc. We propose adaptive techniques that enable a gossip-based protocol to automatically and dynamically adapt itself to the network topology. Our techniques are capable of coping with wireless losses and unpredictable node failures that affect network connectivity over time. The adaptive techniques also allow the sensor network applications to specify a desired reliability for disseminating messages. Nodes automatically adapt their behavior to satisfy such requirements. The resulting protocol is completely decentralized. We present thorough experimental results to evaluate our “Smart Gossip” proposal, and demonstrate its benefits over existing gossip protocols.