A skewed distributed indexing for skewed access patterns on the wireless broadcast

Abstract

Data broadcast is an efficient way to disseminate information to a large number of mobile clients in the wireless environment. Adding an index data organization to the broadcast file can save client power consumption with little increase in client waiting time. The existing index technologies only consider equal access probabilities of data items. However, in real-life applications, some data items may be more popular than others; that is, access patterns of clients are skewed. In this paper, we propose a skewed distributed indexing, SDI, which considers the access probabilities of data items and the replication of index nodes. The proposed algorithm traverses an index tree to determine whether an index node should be replicated by considering the access probability of its child node. In our experimental results, we have shown that our proposed algorithm outperforms the variant-fanout index tree and the distributed indexing.