Method for detecting network node abnormality

Abstract

The utility model discloses an detection method for detecting the abnormal network nodes. The method arranges the link status zone bit in the communication program; as the communication program starts up, the examination thread is established to examine each link in the link information table, so as to check for the network nodes which are correspondent to the link. If the link status zone bit thereof is under running state and the difference between the present time and the link's message sending time is no less than the pre-set value, the link's correspondent network nodes are judged as abnormal. So, the link shall be closed and informed to the upper level; or, the next link shall be examined. The utility model can conduct the judgment before receiving the other side's heart-beat messages or recv function return or send function return, so as to greatly shorten the testing time used for detecting the abnormal conditions of the network nodes. Also, the call loss is reduced.

Description

technical field [0001] The invention relates to the field of intelligent network communication, in particular to a method for detecting abnormality of network nodes. Background technique [0002] Due to complying with the extremely high requirements of current operators for system security and improving the security and reliability of existing network equipment, dual-network and dual-plane intelligent network systems are more and more widely used. Figure 1 is a plane architecture diagram of a dual-network dual-plane intelligent network. As shown in Figure 1, the dual-network dual-plane intelligent network system consists of an Ethernet switch, a service management node, an integrated interface machine, a service management access node, and a router. In a dual-network and dual-plane intelligent network system, each network node is divided into active and standby. Under normal circumstances, the main machine is in the running state, and the standby machine is in the standby st...

AI Technical Summary

Problems solved by technology

[0004] At present, there are three ways to detect abnormalities in the dual-network and dual-plane intelligent network communication program: 1. The communication thread does not receive the heartbeat message from the other party within the agreed time

2. The system call recv failed to return

3. The system call send failed to return

The first method takes at least 25s to detect anomalies. During this time, the upper-layer application thinks that the link is normal and still sends messages, resulting in call loss.

The second and third methods rely on the four-way handshake message at the end of the Transmission Control Protocol (TCP). When waiting for the peer to respond to the close message, the local TCP is in the TIME_WAIT state, which lasts twice as long The longest segment life cycle (MSL), generally between 1 minute and 4 minutes, during this period, recv and send will not return failure, which will also cause call loss

Images