Reduction of the Communication Traffic for Multi-word Searches in DHTs

In large-scale wireless networks, efficient file searching mechanisms are essential for users. Many studies have been done on a distributed hash tables (DHTs) which are one kind of peer-to-peer system. In DHT networks, we can certainly and effectively get the desired files if they exist. However, much communication traffic comes from multi-word searches. If we increase the storage redundancy, we can reduce the traffic. However, storage constrains of mobile devices make doing so impossible. In our previous work, we proposed a divided bloom filter to reduce the traffic in DHTs. However, the effect of divided bloom filters depends strongly on their parameters, and optimizing them is difficult. Hence, we have developed and tested the method to determine optimum parameters of divided bloom filters.