摘要
网络流量特征计算是网络流量分析的一个重要步骤,对于海量网络流量数据,并行化计算网络流量特征是高效网络流量分析的重要方法。针对传统单机处理成本高、可扩展性差的问题,提出一种基于Map Reduce编程模型的网络流量分析方法,并行实现网络分流和流量特征计算。通过使用Hadoop平台对实际数据进行分析,统计常用网络流量属性特征,实验表明,该方法分析网络流量特征的结果准确可信,且适合分析大流量数据。
Internet traffic characteristics computation is an important step of internet traffic analysis, in the face of massive internet traffic data, parallel computing internet traffic characteristics is base to prompt the performance of internet traffic analysis. In order to solve the problems of poor expansibility and high cost that caused by the traditional stand-alone, an internet traffic characteristics analysis method based on MapReduce, which parallel processes diffluent internet traffic and characteristics computing, was proposed. By using the Hadoop platform, the actual data was analyzed and the common internet traffic characteristics were computed. Experiments show that the method is reliable to analyze the characteristic of the internet traffic and it is suitable for analyzing large traffic data.
出处
《电信科学》
北大核心
2014年第12期76-81,共6页
Telecommunications Science
基金
重庆市应用开发计划基金资助项目(No.cstc2013yykf A40006)
关键词
网络分流
特征
并行计算
MapReduce, diffluent internet traffic, characteristic, parallel computing
