Fuzzy Search Controller in Unstructured Mobile Peer-to-Peer Networks

Abstract

Typical mobile P2P protocols largely relied on inflexible techniques such as, flooding, replicating, random walks and selective forwarding to route queries and discover objects of interest, which incur a relatively high search time due to remarkable network traffic, multiple copies and duplication of query messages. Recently proposed P2P protocols are able to reduce search time but due to peer mobility, these protocols cause low hit rate and high overhead. Thus, this article proposes novel fuzzy controller based probabilistic walk for unstructured mobile P2P networks to reduce search time. The search time is reduced by jumping a query walker to 2-hop away selected ultrapeer through fuzzy controller. Furthermore, each ultrapeer share its pong cache with its directly connected ultrapeer to increase hit rate and reduce network overhead. Simulations show that the fuzzy search controller gives better performance than the competing protocols in terms of reducing response time by 10% and increasing hit rate by 15% in different mobility scenarios.