A component-based architecture for adaptive bandwidth allocation in Wireless Sensor Networks

In this paper we investigate the problem of reactive on-line bandwidth allocation in adaptive Wireless Sensor Networks. In adaptive WSN for monitoring applications, nodes can react to physical events by adapting their behavior, for instance increasing the amount of computation to be performed, or reducing the amount of consumed energy. Accordingly, the nodes may increase or decrease their bandwidth requirements. Hence, a reconfiguration algorithm is needed to reallocate the bandwidth to the different nodes, to maximize a certain global quality index. To address this problem, we first present a component-based methodology to model nodes and cluster of nodes as components. Then, we use this methodology to decompose the global optimization problem into simpler local problems, so to enable the design of scalable heuristic algorithms. We propose two different heuristic algorithms, and we compare their performance by extensive simulations against a global optimization procedure based on the Simplex algorithm. We demonstrate that our proposed heuristics perform very well under almost all circumstances.