REEL: A real-time, computationally-efficient, reprogrammable framework for Wireless Sensor Networks

Abstract

Remote reprogrammability and computational-efficiency of sensing nodes are two challenging and open research issues in Wireless Sensor Networks (WSNs). Virtual Machines (VMs) provide a reprogrammable and hardware-independent runtime environment for WSN-based applications but at the expenses of a lower efficiency. On the contrary, code native loaders provide higher computational efficiency but do not support heterogeneous networks and are not able to optimize the code size. In this paper we suggest REEL, a real-time computational-efficient reprogrammable VM-based framework for WSNs. The proposed approach guarantees a higher efficiency compared to traditional VMs and a shorter execution time which, in turn, implies a lower energy consumption. Experimental results on a real application confirm the validity of the proposed solution.