针对幂律泊松模型推测网络蠕虫传播路径  

Reconstruction of worm propagation path for power-law Poisson model

作　　者：石伟[1] 李强[1] 向阳[1] 鞠九滨[1] 

机构地区：[1]吉林大学计算机科学与技术学院,长春130012

出　　处：《东南大学学报（自然科学版）》2008年第A01期77-80,共4页Journal of Southeast University：Natural Science Edition

基　　金：国家自然科学基金资助项目(60703023);吉林省科技发展计划资助项目(20080108)

摘　　要：为了尽早获取网络蠕虫的传播路径,在对Internet流量的幂律泊松分布进行假设检验与参数计算的基础上,提出了幂律泊松流量分布模型下推测网络蠕虫传播路径的k聚积算法.采用数学方法证明了k聚积算法的有效性.通过模拟环境进行实验,研究了参数k对算法准确率的影响,并对算法有效性进行了验证.实验结果表明：当通信流量中入度幂律分布参数γ值大于3,k在0.3-0.5之间时,k聚积算法的准确率最高;当γ值介于2-3之间,k在0.5-0.7之间时,算法准确率最高;当γ值小于2,k在0.7-0.9之间时,算法准确率最高.针对不同的入度幂率分布情况,通过参数k的恰当选择,k聚积算法可以达到89%的准确率.通过试验可以选择参数k在不同幂率分布参数下的最优取值范围,使得k聚积算法对不同的流量分布模型具有较好的适应性.Abstract： In order to ascertain the Intemet worm propagation path as soon as possible, based on the hypothesis testing and parameter calculation of the power-law Poisson distribution in the Intemet traffic, k accumulation algorithm is proposed for reconstructing worm propagation path under power-law Poisson model. The effectiveness of k accumulation algorithm is proved using the mathematical method. Through simulation experiments, the influence of parameter k on the accuracy of the algorithm is investigated and the effectiveness of the algorithm is verified. Experimental results indicate that, if the parameter of power-law distribution γ〉3, the proposed algorithm will achieve a highest accuracy if k is set between 0.3 and 0.5; while in the other situation when yvalues between 2 and 3, a k value of 0.5 to 0.7 will get optimum performance; when γ〈2, the accuracy of the algorithm will be the highest if k is in the range of 0.7 to 0.9. For different power-law distribution situations, k accumulation algorithm achieves a high accuracy of 89% by selecting parameter k appropriately. Range of k in which the proposed algorithm obtain optimum performance in different power-law distributions can be chosen by experiments, so that k accumulation algorithm has a better adaptability in different traffic models.

关 键 词：蠕虫 传播路径 幂律泊松 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]