An Adaptive Caching Technique Using Learning Automata in Disruption Tolerant Networks

In Disruption Tolerant Network(DTN), determining the exact location of data and amount of delay to query the data by a requester is a major concern. It is costly for a node to maintain information of opportunistic paths to every other node in a DTN. Identifying appropriate caching locations is a difficult task. In this paper, we propose a technique for selecting the nodes for caching data based on the past performance of the respective nodes so that data queries can be fulfilled with less delay. In the proposed scheme, initially a Connected Dominating Set(CDS) is selected which serves as virtual backbone for the network. The CDS serves as the initial probable set of caching nodes. Further, an accurate caching set has been determined using a learning method called Learning Automata(LA). A Least Recently Used cache replacement method is used for replacing the data from the buffer of caching nodes, once the buffer is filled. The caching nodes cache the data generated by source node and forwards the data to other caching nodes. A requester node broadcasts the data query to the neighbors and the query is replied by a caching node. Simulation results show the efficacy of the proposed approach in terms of data delivery ratio and packet delay.