A gossip-based energy efficient protocol for robust in-network aggregation in wireless sensor networks

We consider the problem of energy efficient and fault tolerant in-network aggregation for wireless sensor networks (WSNs). We propose a new protocol for in-network aggregation in WSNs, which is energy efficient, achieves high lifetime, and is robust to the changes in the network topology. Our protocol, gossip-based protocol for in-network aggregation (GPIA) is based on the spreading of information via gossip. GPIA is not only adaptive to failures and changes in the network topology, but is also energy efficient, unlike the pure version of gossip. Energy efficiency of GPIA comes from all the nodes being capable of selective message reception and detecting convergence of the aggregation early which makes it significantly better than the pure version of gossip in energy consumption. We experimentally show that GPIA provides significant improvement over some other competitors like the Ridesharing, Synopsis Diffusion and the pure version of gossip. GPIA shows ten fold, five fold and two fold improvement over the pure gossip, synopsis diffusion and Ridesharing protocols in terms of network lifetime, respectively. Further, GPIA retains gossip's robustness to failures and improves upon the accuracy of synopsis diffusion and Ridesharing.