Consistent data sharing in mobile ad hoc networks

In Mobile Adhoc Networks (MANETs), data availability is lower in comparison to the conventional fixed networks because of their mobility and resource-constrained characteristics. Replication and caching have been adapted to improve the performance in these constrained environments. These techniques significantly improve the efficiency of information access in MANETs resulting in reduction of the access latency and bandwidth usage. The paper proposes a cluster-based caching and data consistency algorithm where the clustering is done employing learning automata to support stability. The caching related information is maintained by the clusters members and the cluster heads. Apart from the avoidance of cached data redundancy and the reduction in the average access latency, the proposed scheme also enables TTL-based cache invalidation to avoid the stale data availability. The network is simulated using NS-2 to evaluate the performance of the proposed algorithm. The simulation results show an improvement in the cache hit ratio and reduction in average latency compared with few existing caching schemes.