A cloud-assisted malware detection and suppression framework for wireless multimedia system in IoT based on dynamic differential game

As a novel dynamic network service infrastructure, Internet of Things (IoT) has gained remarkable popularity with obvious superiorities in the interoperability and real-time communication. Despite of the convenience in collecting information to provide the decision basis for the users, the vulnerability of embedded sensor nodes in multimedia devices makes the malware propagation a growing serious problem, which would harm the security of devices and their users financially and physically in wireless multimedia system (WMS). Therefore, many researches related to the malware propagation and suppression have been proposed to protect the topology and system security of wireless multimedia network. In these studies, the epidemic model is of great significance to the analysis of malware propagation. Considering the cloud and state transition of sensor nodes, a cloud-assisted model for malware detection and the dynamic differential game against malware propagation are proposed in this paper. Firstly, a SVM based malware detection model is constructed with the data sharing at the security platform in the cloud. Then the number of malware-infected nodes with physical infectivity to susceptible nodes is calculated precisely based on the attributes of WMS transmission. Then the state transition among WMS devices is defined by the modified epidemic model. Furthermore, a dynamic differential game and target cost function are successively derived for the Nash equilibrium between malware and WMS system. On this basis, a saddle-point malware detection and suppression algorithm is presented depending on the modified epidemic model and the computation of optimal strategies. Numerical results and comparisons show that the proposed algorithm can increase the utility of WMS efficiently and effectively.