基于SNMP故障管理中的告警振荡和抵消方法

针对网络拥塞或设备处理器忙时造成网络管理系统对设备运行的故障状态产生误判,从而导致网络管理系统产生告警振荡的原因及其统计特性进行了深入研究,提出了一种基于概率统计的投票算法来有效地消除告警振荡.该算法通过投票表决来判断是否出现告警振荡,并能抵消在指定门限时间范围内出现的告警振荡,从而有效地消除虚假的告警通知并提高网络管理的效率.在实际的公共接入无线局域网网络管理系统中具体实现了该算法,实验结果证明了该算法的有效性.

Reduction of Alarm Vibration in SNMP-Based Fault Management

When congestion occurs in a managed network or a managed device is too busy, an NMS (network management system) may acquire an improper state of the managed device, therefore results in alarm vibration in the NMS. By the analysis of the causation and statistic nature of alarm vibration, a voting algorithm based on probability model is proposed to reduce alarm vibration and improve the efficiency of management. The algorithm takes a vote on the results of continuous detection of a managed device's state to determine whether or not an alarm vibration has arisen in the NMS and eliminates the vibration occurred during the time range specified by a threshold. The validity of the proposed algorithm has been proved in actual public access wireless local area network management system.