Study on Detection Algorithm of DDoS Attack for Cloud Computing

In order to solve the problem of distributed denial of service (DDoS) attack for cloud computing, a DDoS attack detection algorithm was proposed based on feature analysis and Kalman filter. According to the difference of the frequency of accessing the cloud servers between ordinary users and distributed denial of service attacks user, the behavioral features were used as the detection objects. The number of accessing the cloud server and the accessing behavior features computed by IP number were counted within a certain time. The linear prediction algorithm was used to predict the behavior features. Kalman filter was used to correct the prediction value. The corrected value and the prediction value were compared with each other. The value which exceeds the specified threshold was judged as the DDoS attack. Simulation result shows that the algorithm has better detection rate and lower false detection rate. It has good application value in network security defense.