Broadcast Search in Power-Law Networks

Broadcast search strategy is widely used and plays a key role in unstructured Peer-to-Peer applications, many of which exhibit a power-law like degree distribution. Understanding and improving scalability and efficiency of broadcast search in power-law networks can give huge benefits to these systems´ search performance. In this paper, we analyze the impacts of topology structure and peer degree´s heterogeneity. Then, we propose a broadcast search strategy combined with index replication technology and evaluate its search performance in power-law networks. The experiment shows broadcast search with index replication can achieve much better search performance in both reducing peers´workloads and improving search success rate.