A Storage Centric Approach to Scalable Sensor Networks

With the advent of Machine to Machine (M2M) communication, we are witnessing an increased interest towards technologies that will enable efficient and reliable operation of Wireless Sensor Networks (WSN). Such networks are expected to include a large number of sensor devices which will generate large body of M2M traffic. To reduce the impact of this M2M traffic, efficient storage and retrieval methods should be employed in a distributed manner for the successful deployment of this technology. In this paper, a distributed storage solution is presented with the aim of reducing the impact of M2M traffic on data centres and the network backbone. The reliable and efficient storage of the sensor data is established by taking advantage of codes with Maximum Distance Separable (MDS) properties. The solution is implanted in Contiki OS using RPL protocol [1] and its performance is evaluated through simulations. Furthermore, to realise such a system, a centralised clique finding algorithm is demonstrated and benchmarked against a solution that uses a brute force approach.