基于优先权的分布式网络设备智能轮询策略研究

针对大型网络其地域分布广,监测对象较多,监测周期过长,实时性较差且管理流量负载较大等问题,提出一种分布式网络设备智能轮询策略。在域内,针对轮询周期调整问题,通过构造体现管理者意志的评判矩阵和求解反映数据特征的DFT(离散傅立叶变换)确定轮询优先级和轮询周期取值范围,提出了基于变化率的轮询周期调整算法;在域间,通过触发和查询实现对特殊事件的采集。最后进行了仿真,表明该策略能够在减轻网络负载的同时,准确刻画网络的状态。

Study on Distributed Intelligent Polling Strategy for Network Devices Based on Priority

For the characteristics of long monitoring period,poor real-time behavior and heavy traffic load of the large network,which are caused by the wide geographical distribution and multiple monitoring objects,a novel distributed intelligent polling strategy is proposed. In the region,a method to calculate the priority and the range of the polling period based on the manager will and the DFT (Discrete Fourier Transform)is presented. What’s more,a polling period adjusting algorithm according to the change rate is illustrated in detailed. Between the region,the information is gathered by triggering and queries for special events. Finally,a simulation analysis is given out to prove the efficiency and accuracy.