The Optimization Of P2P Overlay

As a type of new distribute system, P2P system have became one of the most popular Internet distribute systems because of the out feature of peer-to-peer computing on managing and using distribute information and compute resource. Specially, decentralized unstructured peer-to-peer systems have applied widely in Internet, because their simple and convenience. But the resource location is very difficult in those systems because of the instability of the system's topology and the random of the resource's location. So currently they use flooding-based search, which lead to a large traffic overhead on network. And the topology mismatch between physical network and P2P overlay is also a serious problem. It limits the capability of resource location technologies, especially for the flooding-based search.There are a lot of researches on the optimization of P2P overlay in currently, and the results of those works are used to improve the performance of resource location and message routing technology. In this paper, we analyzed the characters of topology mismatch firstly, and obtained a character that can judge the node sequence mismatch locally. Then, we propose a new optimization method based on the character. The method is named HAND and it has several advantages comparing with previous ones. We did comprehensive simulation and analysis, and proved that HAND is convergent, and it is very fit for the dynamic distribute environment.This method has several advantages comparing with previous ones. Firstly, it is a completely distribute algorithm and especially works in large-scale P2P systems. Secondly, its optimization overhead has decreased over 50 percent that compared with LTM, and the performance of optimization improved 10 percent to 20 percent.