Iterative P2P credit management scheme based on clustering

Abstract

The invention provides an iterative P2P credit management scheme based on clustering. The scheme is characterized by: based on a traditional iterative P2P credit management scheme, introducing a clustering method, calculating a feedback polymerization degree through the clustering and combining a global reputation value of a recommended node so as to take as a recommended weight; using a clustering process to logically divide a whole network into several small-scale networks so that iterative processing is performed in the small-scale networks and a iterative convergence speed can be accelerated; finally, combining the global reputation value of the node according to a size of the network scale; during service selection, firstly ordering the candidate service nodes according to the credit value; selecting in probability a candidate subset which ranks higher. The service nodes which possess the higher credit value have greater probability to be selected. Therefore, the service nodes can be guaranteed to be generated in the candidate set with the higher credit. Simultaneously, the service node with the highest credit value can be avoided to be selected all the time and overload of the node can be avoided.

Description

technical field [0001] The invention relates to the technical field of P2P, in particular to a cluster-based iterative P2P reputation management scheme. Background technique [0002] P2P technology can support nodes to share resources in a large-scale distributed environment, so it has become a practical technology that has been extensively studied and developed rapidly in recent years. P2P computing architecture has played an important role in different application fields such as peer-to-peer collaboration, resource sharing, and knowledge management. The most widely used is based on the sharing and downloading of resource files, such as BT, PPLive and other practical software have a growing user base. The P2P network distributes the processing response of the central server of the traditional C / S network to each node in the network. Weaken or cancel the status of the central server and share it with all nodes in the network. Resources and services are scattered in the net...

AI Technical Summary

Problems solved by technology

[0008] The problem solved by the present invention is that the traditional iterative P2P reputation management scheme cannot rationally allocate recommendation weights when calculating reputations, resulting in inaccurate calculated reputation results; the iterative solution process carried out in the entire network is slow to converge, thus making the solution more efficient Low; the service node always chooses the node with the highest reputation, causing a single point of overload

Images