Popularity Biased Hybrid Search in P2P Systems

Hybrid search, which combines the flooding and DHT techniques, has been employed for efficient content-locating in peer-to-peer (P2P) systems. Choosing a flooding or a DHT subjects to the population of desired data. However, existing studies summarizes the popularity only based on local information, or does not take into account the dynamic factor of P2P systems. Thus, the users suffer from long search latency when performing hybrid search due to inaccurate estimations of the resource popularity. In this paper, we propose a popularity biased hybrid search method in systems. Our model allows the P2P systems dynamically detect the content popularity, and accurately make a decision of search methods. In this way, our work effectively reduces the query traffic cost and response time. The result our trace-driven simulations report shows that our model outperforms existing approaches in search efficiency