An efficient resiliency scheme for data centric storage in wireless sensor networks

Abstract

In this paper, we consider the problem of providing efficient resiliency for data centric storage in wireless sensor networks, which is recently gaining interest. Three types of canonical storage methods are available in the literature for storing the sensed data, namely, external storage, local storage, and data centric storage. In these methods, a geographic hash table achieves data centric storage where each data item is stored at a corresponding hashed location. In this paper, we introduce a novel concept, called multiple hash locations for storing sensed data. In the proposed method, by considering the real time nature of the sensing data, we provide various resiliency levels for various kinds of data. This introduces different resiliency levels for different data items based on their importance. Earlier schemes, available in the literature, do not consider this variance and provide resiliency common to all data types. These schemes also suffer from single point of failure such as central node failure, and also from clustered node failures. Where as, our scheme is fault tolerant for central node failures and clustered failures. Our scheme also avoids the communication caused by local replication along the perimeter of hashing location by introducing geometric shape concept. Each sensor node in this geometric area is aware of the actual replicating node for a specific data item, known as node awareness. The size of the geometric shape depends on the density of sensor nodes in that area. In our scheme, routing of data is divided into two parts, first, from the origin to any node in the geometric area, and second, from any node to actual destination inside the geometric area. Since, hashing location is now represented by a geometric shape, the proposed scheme does not actually dependent on the underlying routing algorithm. © 2004 IEEE.