Threat monitoring method for virtual network insider based on behavior traceability

Conference: CAIBDA 2022 - 2nd International Conference on Artificial Intelligence, Big Data and Algorithms
06/17/2022 - 06/19/2022 at Nanjing, China

Proceedings: CAIBDA 2022

Pages: 5Language: englishTyp: PDF

Authors:
Li, Haosheng; Bao, Jian; Yun, Kai (State Grid XinJiang Information & Telecommunication Company Urumqi Xinjiang, China)

Abstract:
In order to improve the monitoring accuracy of internal threats in the virtual network and reduce the content consumption during network operation. This paper proposes a behavioral traceability-based threat detection method for virtual network insiders. Based on the threat characteristics inside the virtual network, we use the long-short-term memory network to build an accurate user behavior model. In this paper, the variational autoencoder is used to optimize the data association and complete the traceability of user behavior. And on this basis, it uses the long short-term memory network to describe the hidden state of virtual network information, and realizes the monitoring of internal threats in the virtual network. Experiments show that the recognition accuracy rate of this method for different management operations is higher than 95%, and the growth rate of memory usage is lower than 7%, which effectively improves the accuracy of threat monitoring inside virtual networks.