端到端冗余流量消除技术的指纹选择算法研究

为了兼顾服务器端的内存消耗和有效性,在分析了现有端到端冗余流量消除技术中的指纹选择算法后,提出了贪婪指纹选择算法。该算法认为两个数据包中相同片段的邻接片段很有可能也是相同的,基于这一原理,邻接片段的指纹也将被选为代表指纹,有效地避免了在服务器端缓存整个数据包,大大减少了服务器端的内存消耗。仿真实验结果表明,该算法与现有的指纹选择算法相比,能够大大减少服务器端和用户终端的内存消耗,同时冗余消除率在10%-50%之间,与现有算法相近,证明了该算法的有效性。

Fingerprinting algorithm of end-to-end redundancy elimination in network traffic

To take into account both memory consumption and efficiency,existing fingerprinting algorithm in end-to-end redundancy elimination is analyzed and a new greedy fingerprint algorithm is proposed.The algorithm is established on the theory that the two seg-ments which adjoin to two same segments in two different packages are more likely to be same.Based on the hypothesis,the fingerprint of the neighboring segment should also be chosen as representative fingerprint.The algorithm avoids caching the whole package...