Evaluating effects of data popularity on performance of caching strategy in pull-based data boradcast systems

Abstract

A data broadcast in wireless environments is one of efficient methods to disseminate data from a server to a number of clients. In data broadcast systems, a caching can be a useful technique to alleviate performance degradation by reducing contention for narrow bandwidth due to broadcast requests. In this paper, for pull-based data broadcast systems we propose a caching strategy which is expected to have prefetching effects by considering data popularity among clients. Then, we evaluate the proposed strategy through a simulation approach to analyze the prefetching effects. According to the results, in normal case the proposed strategy generally shows better performance than other strategies in terms of average response time.