Study On Isp-aided P2P Resource Management Mechanism

In the past few years, P2P systems scale have been expanding, providing with large number of users, widely distributed. Meanwhile, the various aspects of resource management in P2P systems have becoming increasingly appearance, and impeding P2P applications.Currently, P2P overlay network topology construct on the underlying physical network and most of them does not take into account the actual situation of physical network, resulting topology mismatch problem. It leads to inefficient use of the underlying link and upper application management mechanism. Unstructured P2P system`s topology mismatch problem is particularly serious as lack of central management mechanism. To solve this problem, we propose the Unstructured P2P system topology optimization base on the physical location information of node. It rely on the ISP and P2P system mutual cooperation. ISP feedback the global and local location of nodes, node used this information as a criteria to guide their behavior of choice neighbors. Thus making the system nodes connected nearby, reducing the topology mismatch problem. Simulation results show that the optimization strategy can reduce the system topology mismatch problem and traffic caused by P2P system impact on the underlying link. When the resources is rich in system, it can significantly reduce the search cost and improve distribution performance of P2P systems.Prior planning ISP traditional traffic engineering to improve the efficiency of network transmission, based on historical flow of the sample collection and access between the end of AS business model, optimize network traffic and traffic load balancing. The traditional Internet business, this method enables the transmission efficiency of the network is close to optimal. However, it requires application-level fluctuations in traffic demand trend forecasts based on sampling within the set P2P systems are generally adopting selfish resource management strategies, although improved the overall performance and user experience, produced a great traffic and aggravated the backbone network's congestion, puzzled the Internet Service Provider(ISP). ISP take many measures such as speed limit, block, etc. There measures greatly undermined the long-term development of P2P systems. Address this problem, we Proposed the ISP-P2P cooperation traffic control strategy based on local information. It took into account ISP traffic management objectives and the performance of P2P systems. ISP base on their understanding of the underlying network topology, select a number of key links as the monitoring link and feedback the nodes of these link usage. Nodes use the information to guide their neighbors selection, adjust the sending or receiving ratio between neighbors at application layer: reduce the ratio which through overload link, and increase ratio through the others. Through this strategy, to guide the traffic flow to the light load and balance the traffic, while overcoming the disadvantage that slow traffic adjustment of traditional traffic engineering. For P2P system, by avoiding the traffic through some congestion links, reducing net latency due to congestion. While picking better performance connection for the transmission between neighbors and corresponding increase date ratio in these connections, achieving a flow shifting effect and improve the ratio of data distribution among nodes, improving the distribution of overall system. The final simulation results also show that strong.