论文部分内容阅读
根据云计算数据中心网络(data center networks,DCNS)架构的特点,从网络架构的角度对低速率拒绝服务(low-rate denial of service,LDo S)攻击进行建模。提出基于可用带宽欧氏距离的LDo S攻击检测方法,其本质是依据LDo S攻击导致同一路由域内所有链路可用带宽同时增大的特征,将可用带宽的平均欧氏距离作为LDo S攻击检测测度。改进了传统的探测间隔模型(probe gap model,PGM),并将其专门用于云计算环境下的可用带宽测量。在实际的网络环境中对LDo S攻击效果和LDo S检测性能进行测试,结果表明:1)DCNS内的LDo S攻击比洪水式拒绝服务(flooding denial of service,FDo S)攻击更具危害;2)所提出的检测方法能够准确检测LDo S攻击,检测率达到98%。
Low-rate denial of service (LDoS) attacks are modeled from a network architecture perspective, based on the characteristics of the DCNS (data center networks) architecture. The LDoS attack detection method based on available bandwidth Euclidean distance is proposed. The essence of this method is that the LDoS attack leads to the increase of available bandwidth of all the links in the same routing domain at the same time. The average Euclidean distance of available bandwidth is taken as LDoS attack detection measure . The traditional probe gap model (PGM) has been improved and used exclusively to measure the available bandwidth in a cloud computing environment. The results of LDo S attacks and LDo S tests show that: 1) LDoS attacks within DCNS are more harmful than flooding denial of service (FDoS) attacks; 2 ) The proposed detection method can accurately detect LDo S attacks, the detection rate of 98%.