Policy-Based Reprogramming for Wireless Sensor Networks

Program codes running on the sensor nodes need to be updated from time to time to fix bugs, tune the parameters, insert new functions, or delete useless codes. In this paper, we propose a policy-based reprogramming framework that can optimize the reprogramming process based on the execution characteristics of modules that constitute the WSN applications. To accomplish this we model the characteristics of a module as a tuple of the following four criteria; region, residence, synchronization and real-time properties of the module. According to the class of a module and the type of a reprogramming operation, we adaptively use 3 different reprogramming techniques; efficient code dissemination using connected dominating set, selective code dissemination, and code acquisition. Experimentation results confirm that this policy-based reprogramming can achieve substantial improvement in both the energy consumption and the reprogramming latency compared to existing solutions for various reprogramming scenarios.