Approach of DDoS Attacks Prediction and Detection with HEQPSO-SVM Algorithm Based on Date Center Network

The paper presents a novel quantum particle swarm optimization algorithm which is based on hybrid entropies( HEQPSO algorithm) and support vector machine( SVM) algorithm applying to the network of data center on the issues of predicting and detecting distributed denial of service( DDoS) attacks. The proposed approach collects the network flow at detection server and makes presort for data samples which have the same characteristics. The evolution speed factor( ESF) and aggregation degree factor( ADF) is introduced into the HEQPSO algorithm to optimize the error penalty factor C and the kernel width of Gaussian radial basis function( RBF) σ of SVM classifier. Meanwhile,the hybrid entropies strategy is designed to identify the potential DDoS attacks by the linear equation of sample characteristics. Simulation experiment demonstrates that the proposed method remarkably predicts and detects the intrusion of DDoS attacks. Moreover,HEQPSO algorithm has better generalization ability and smaller error performance in comparison with the traditional QPSO algorithm and DCWQPSO algorithm in terms of algorithm execution time,average iterations,average relative variance( ARV) and root mean square error( RMSE). Compared to the four classic DDoS prediction with the proposed detection method in response time,detected ratio,mis-detected ratio and accuracy,it is proven that the proposed method is superior.