Characterizing the Multicast Nature of Malicious Flows in CCN via SIS Epidemic Model

Content-centric networking (CCN) should provide enhanced protection over current network. However, it is still threatened by malicious attackers. This article analyzes the influence of CCN's multicast nature on malicious spread via multicast-based epidemic model. Our multicast-based model well captures the spread dynamic of malicious flow in CCN; it is capable of evaluating node's probability of state transitions. First, we propose a multicast-based model by advancing N-intertwined model with multicast matrix that encodes multicast groups and traffic distribution. We derive possible tight bounds for epidemic threshold. The monotonicity result of this threshold gives us insights into studying impacts of forwarding, caching, and traffic control strategy of CCN on epidemic spread. The optimal control problem is presented to jointly minimize the cost of epidemic spread and immunization. The existence of optimal control is proven. We provide necessary and sufficient conditions for optimality. Based on optimal immunization, we discuss the design of countermeasure in terms of traffic control and forwarding. We propose a targeted immunization strategy to select targeted nodes to deploy detection and countermeasure. Finally, we verify the accuracy of the multicast-based model and study the performance of targeted immunization by simulations. Results show that our model accurately describes malicious spreading, and targeted immunization reduces infected fraction with low cost.