Internet of Things (IoT) is a novel Internet revolution and it creates the Objects themselves communicate information, recognizable; attain intelligence, regarding themselves as well as they can access information, which is collected by other things. Nevertheless, the IoT network of physical devices and objects is frequently vulnerable to attacks such as Denial of Service (DoS) and DDoS. Till now, numerous studies have been conducted to eradicate the aforesaid threat problem and in this case, this work tries to present a novel attack detection model. The adopted attack detection model creates DevOps interconnection as it generates connections among expansion as well as IT operations. Moreover, a developed attack detection system assures the operations security of diverse applications. Therefore, the adopted model involves two main stages such as adopted feature extraction well as classification. From each application, the data are processed in the primary phase of the feature extraction, wherein statistical as well as the higher-order statistical features are merged. Then, to classification procedure, the extracted features are fed; here it ascertains the occurrence of the attacks. This work attempts to use the optimized Deep Belief Network (DBN) model for the classification procedure, in that the activation function is optimally tuned. In addition, optimal tuning procedure is a serious feature and is obtained merely using some optimization logic. This work develops a novel hybrid approach called GSA-PSO, which is the combination of Genetic Algorithm (GA), Simulated Annealing (SA), and Particle Swarm Optimization (PSO) correspondingly. At last, the adopted model performance is evaluated with the existing model regarding certain performance metrics.
Buy article PDF