Detection of Distributed Denial of Service Prevention (DDoSP)
There are many solution based methods created against Distributed Denial of Service (DDoS) attacks are focused on the Transmission Control Protocol and Internet Protocol layers as a substitute of the high layer. An extended hidden semi-Markov model is proposed to describe the browsing habits of web searchers. A forward algorithm is derived for the online implementation of the model based on the M-algorithm in order to reduce the computational amount introduced by the model’s large state space. Entropy of the user’s HTTP request sequence accurate to the replica is used as a principle to measure the user’s normality. Finally, experiments are conducted to validate the authors’ model and algorithm.