基于布隆过滤的藏区Web站点流量识别

摘要

针对被管网大流量条件下藏区Web站点流识别算法准确性低、鲁棒性差等问题,提出一种基于布隆过滤BF(bloom filter)的藏区Web站点流量识别方法.给出能够描述藏区Web站点流的特征字段,形成关键字集合,并映射为BF中的位数组;在给定假阳率的情况下,利用Hash函数对被管网中数据包的相应特征字段进行Hash映射操作,识别该包是否为藏区Web站点流量的网络包.实验结果表明,该方法呈现了较高的准确性,识别率保持在92.3％以上,在网络流量较大时仍表现出较强的鲁棒性.%Aiming at the problem of poor accuracy and robustness of current Tibetan Website flow identification algorithm under the condition of high flow rate of the monitored network,a Tibetan Website traffic recognition method based on Bloom filter (TWTBF) was proposed.The packet feature fields of Tibetan Website traffic were extracted to form a keyword set which were then mapped to bit array of Bloom filter (BF).In the case of given some false positive rates,the corresponding feature fields of packet in monitored network were computed and mapped using Hash function,which decided whether or not this packet belonged to Tibetan Website traffic.Experimental results show that,the proposed method not only shows high accuracy with a correct identification rate of 92.3％ above,but also presents strong robustness under the heavy network traffic.