SEIR-based botnet propagation model

The botnet in real network environments has scale-free characteristics, there are severe uneven distribution characteristics among nodes. Therefore, the botnet host propagation model based on random network is unrealistic, and feature nodes and random nodes must be added in the modeling process to reflect the generic scale-free network characteristics. In this paper, the traditional SIR model is improved by analyzing the traffic of different types of botnets and introducing the latency period as a variable quantity, marked as Exposed (E), which can more accurately reflect the propagation characteristics of botnet hosts in real networks. Based on the improved SEIR model, a control strategy aimed at scale-free botnet is proposed, which provides a new theoretical basis for botnet detection, tracking and containment.