Support-Based Prefetching Technique for Hierarchical Collaborative Caching Algorithm to Improve the Performance of a Distributed File System

Prefetching and client side caching are the two important techniques used in distributed file systems (DFSs) for improving performance. Collaborative client side caching techniques are discussed in the literature for reducing the read access time in the DFS. In this paper, we have proposed to have a new type of collaborative caching algorithm which includes not only local caches but also a global cache present in the DFS. We have also proposed a novel support-based technique for prefetching the blocks of the popular files. By combining support-based prefetching with the hierarchical collaborative caching, we have proved through simulation that the read access time can be reduced in the DFS. For performance comparison purpose, we have considered the default read algorithm of Hadoop DFS and the collaborative caching algorithm proposed in the literature.