摘要
为保证系统在给定故障管理带宽阈值下,仍然具有较高的故障征兆信息采集频率,该文提出了一种故障管理智能轮询算法。该算法根据故障征兆参数值的时间属性,使用离散傅立叶变换将参数值分解为不同频率的正弦函数,进而根据系统预先设定的各参数故障告警阈值,决定使用分解得到的最大或最小频率作为轮询间隔时间。最后,通过试验证明该算法具有较高的效率和可靠性。
In order to keep the system with higher collection frequency of fault symptom information in initial bandwidth threshold of fault management, an intelligent polling algorithm is presented. This algorithm decomposes parameter value which includes the time variations of fault symptom into sinusoids with different frequencies by means of discrete Fourier transformation. According to various fault alert initial threshold, the max or rain frequency from decomposition is regarded as polling interval time. This algorithm is proved with more efficiency and reliability than others in experiment.
出处
《计算机工程》
CAS
CSCD
北大核心
2007年第11期126-128,共3页
Computer Engineering
关键词
简单网络管理协议
故障管理
智能轮询算法
Simple network management protocol（SNMP）
Fault management
Intelligent polling algrithm
