Communication/computation trade-offs in wireless sensor networks: Comparing network-level and node-level strategies

In wireless sensor networks, nodes having limited battery resources convey data to an energy-unconstrained base station. The amount of data transmitted by a node usually depends on how much local processing is performed. In other words, more computation on a node means less communication with the base station and vice versa. Hence improving energy efficiency and prolonging the network lifetime requires a careful trade-off analysis. This analysis may be performed at a network-level or at a node-level. The latter is more fine-grained allowing different nodes to implement different solutions. In our work, we propose a novel mixed integer programming framework to model and optimize node-level strategies. Using this framework, we show that hybrid use of digital signature algorithms in a network could extend the lifetime up to 21.25% as compared to a network-level optimal strategy where all nodes use a single algorithm.