Parallel Routing Method in Churn Tolerated Resource Discovery

Abstract

Structured P2P network has good scalability, robustness and routing algorithm with high performance and low cost. Thus, in theory, structured P2P network is a good choice for network resource discovery. However, because of churn, it has bad performance and can't be used extensively in real world. To overcome this problem, a parallel routing method is presented. Network resource information is organized by the node group. The query is moved in node group by Random Walk, and among different groups by the routing algorithm of FreePastry. The parallel algorithm is proposed and its complexity is analyzed. The experiment results verify that the parallel routing algorithm cost is the same as FreePastry and the traffic load is a little bigger than that of FreePastry, but smaller than that of the other ones. The parallel method is tolerated to churn and is helpful for resource discovery of structured P2P network with churn.