发明名称 NETWORK ABNORMALITY DETECTING METHOD
摘要 <p><P>PROBLEM TO BE SOLVED: To solve the problem that a traffic pattern includes usual traffics other than DoS (Denial of Service) attack traffics and the usual traffics work as a noise in evaluating the similarity of a pattern shape, resulting in a cause of deterioration in tracking accuracy in a method of tracking the DoS attack by using the traffic pattern being a temporal change in the number of packet. <P>SOLUTION: Every time multiple analysis is repeated by wavelet transform, a wavelet coefficient (value indicating amplitude of a low frequency component) is reduced in the usual traffic, and the usual traffic has such a characteristic that the high frequency component has a large amplitude. Accordingly, a filtering technique for eliminating the high frequency component from the traffic pattern is used, thereby eliminating usual traffic being a noise component from the traffic pattern including DoS attack, and thus, highly accurate tracking can be realized. <P>COPYRIGHT: (C)2006,JPO&NCIPI</p>
申请公布号 JP2006246326(A) 申请公布日期 2006.09.14
申请号 JP20050062234 申请日期 2005.03.07
申请人 TOHOKU UNIV 发明人 IZUMI YUJI;NEMOTO YOSHIAKI
分类号 H04L12/66;H04L12/70 主分类号 H04L12/66
代理机构 代理人
主权项
地址