Measuring and modelling the energy cost of reconfiguration in sensor networks [forthcoming]

As Wireless Sensor Networks (WSN) must operate for long periods on a limited power budget, estimating the energy cost of software operations is critical. Contemporary reconfiguration approaches for WSN allow for software evolution at various granularities; from reflashing of a complete software image, through replacement of complete applications, to the reconfiguration of individual software components. This paper contributes a generic model for measuring and modelling the energy cost of reconfiguration in WSN. We validate that this model is accurate in the face of different hardware platforms, software stacks and software encapsulation approaches. We have embedded this model in the LooCI middleware, resulting in the first energy aware reconfigurable component model for sensor networks. We evaluate our approach using two real-world WSN applications and demonstrate that our model predicts the energy cost of reconfiguration with 93% accuracy. Using this model we demonstrate that selecting the most appropriate software modularisation approach is key to minimising energy consumption.