摘要： 为提高网络性能管理与故障管理的实时性和效率，研究信息收集轮询策略，提出一种基于马尔可夫链的智能轮询算法。该算法能根据被监视对象参数值的变化，动态调整轮询间隔时间。实验结果表明，与传统等时间轮询算法相比，该算法可以更准确地刻画数据变化特征，降低数据采集失真度并减小网络带宽，其效率和可靠性较高。

关键词: 网络管理, 智能轮询, 性能监视

Abstract: In order to improve the real-time operability and efficiency of network performance management and fault management, this paper studies the information collection polling strategy and proposes an intelligent polling algorithm based on Markov chain. This algorithm can dynamically adjust the intervals according to the parameters of equipment. Experimental results show that compared with traditional equal-time polling algorithm, the algorithm can describe the data change feature more exactly while decreasing the collection distortion degree and network bandwidth, and has higher efficiency and reliability.

Key words: network management, intelligent polling, performance monitoring

中图分类号: 

• TP393.07