Regular detection system of P2P node survivability and method

Abstract

The invention provides a regular detection system of P2P node survivability and a method. The system comprises an information collecting module, an information processing module, a detecting module and a responding module, wherein the information collecting module collects the node connection information in the recent detection period, and initializes the node status attribute of a routing table;the information processing module determines which nodes need to be detected in the next detection period and which nodes need to be delayed to be detected according to the node status attribute of the routing table; the detecting module detects the nodes which need detection when the next detection period comes, changes all node flag bits in all connection tables, continues to collect returned connection information and updates the node status attribute of the routing table; and the responding module responds correspondingly after the node receives detection information so as to update othernodes. The invention fully takes the communication between the nodes on the network into consideration, counts the prior connection information and determines whether detection is carried out or not.The detection times are reduced, and the network bandwidth is saved.

Description

technical field [0001] The invention relates to the technical field of network measurement and control. Specifically, it is a P2P (peer-to-peer network) node survivability periodical detection system and method. Background technique [0002] The scale of the P2P network is gradually expanding, and the number of nodes is increasing, but the joining and exiting of nodes in the P2P network is uncertain. In this way, in order to ensure whether the nodes in the node's own routing table are alive, it is necessary to detect these nodes. [0003] In Pastry (a point-to-point network routing algorithm proposed by Microsoft and RICE University) and many other routing algorithms, the regular detection method is used to determine whether the nodes in the routing table are alive, that is, each node explicitly detects every time Look at all the node pointers in its own routing table to see if the corresponding node is still in the system, and determine whether the node is alive according...

AI Technical Summary

Problems solved by technology

But the disadvantages are also obvious. This method similar to multicast wastes bandwidth more obviously, especially when the number of node routing table entries is large, the number of nodes to be detected is large.

In addition, this method does not take into account the fact that the nodes that have just communicated are still alive, and often send probe requests to nodes that have just communicated.

Images