Research Of P2P Download Flow Optimization Based On The ISP Active Participation

As the development of the Internet technology, user participation is increasing and all kinds of P2P(peer-to-peer network) software are also widely used gradually, which promotes the development of P2P networks. In a variety of P2P applications, the users are concerned about the resources of the query and download speed mainly, besides, the download speed of some P2P download software is the main target related to the scale of customers group .A lot of bandwidth is being used, because the relative works of ISP is also based on P2P technology , ISP cannot block all P2P applications, inconsistency between P2P network topology and the bearing network topology causes that the ISP cannot monitor increasingly large P2P data flow effectively , making the ISP satisfy online experience of the user by increasing the Internet bandwidth, but using the physical method to improve the network traffic transmission speed is limited on condition, therefore, under limited circumstances of performance of physical devices in the network ,effective management of the network traffic is reasonable .Currently ,the P4P(Proactive network Provider Participation for P2P) program of American Yale University and Washington University is an effective way of a P2P network traffic optimization , aimed at strengthening communication of the ISP and P2P node terminal program and improving the transmission performance of P2P traffic.This paper combines the P2P network traffic optimization research and practice at home and abroad, presents the optimal selection of the target node aimed at improving the download speed of the node resources, as the following:①Add properties of nodes in order to simulate the active involvement of ISP to reflect the region of nodes, fully considering the regional property of the node in the selection of the target node to download. When other parameters of selecting target nodes varies slightly, consider in priority the nodes belonging to same ISP internal or among ISP with minimum communication cost to precede resource download task.②In order to improve download speeds of nodes, we choose the best available resources to download the target node to communicate. The best access to the target is the node query algorithm we improved ,adding the reliability of the target node, ISP property, the remaining bandwidth and the number of hops between the source node and other factors to return a reference resource to download comprehensive ranking of nodes, the top-ranking node is defined as the best available resources downloading node, nodes resources downloading in the P2P internal network uses a distributed downloading way, so that some of the goals nodes in real-time network randomly leave the network has little influence on resources downloading and speed of the download .③In the practice session, we use the unstructured P2P network Gnutella to analyze node resource search downloading process, through the Linux system simulation platform to build P2P networks , integrate Gnutella protocol and create a new large network topology structure in NS-2 network simulator platform, which has an important significance on the understanding of node information and P2P search algorithms, and on this basis, we put forward a set of optimization strategies to download ,fully considering the node selection of region and download links to choose and improving the resources downloading speed of the node.