Adaptive Quality-Aware Replication in Wireless Sensor Networks

Abstract

Typical sensor network deployments are usually built for long-term usage. Additionally, the sensor nodes are often exposed to harsh environmental influences. Due to these constraints, it is mandatory for applications to be able to compensate the failure of nodes. Providing a persistent storage even in the presence of failing nodes demands for replication within the sensor network. However, recent work in the field of replication in sensor networks often does not consider the suitability of the sensor nodes to store replicas in terms of e.g. available storage, energy or connectivity. In this paper, we envision an adaptive quality-aware replication scheme which enables the storage of replicas based on a scoring system reflecting the suitability of a replica node. Furthermore, we propose an adaptable data migration strategy using a weighting function to achieve an adequate placement for the replicas. A resilient storage strategy enables the survival of replicas after migration despite unpredictable node failures. We expect that our replication scheme highly increases the availability of sensor network data despite of node failures and network partitioning requiring only a small number of replicas within the network.