An Active Security Defense Strategy for Microservices based on Deep Reinforcement Learning

Abstract

For container cloud and microservice architectures, the attack process is more complex and diverse. How to improve the defense effectiveness of the system while ensuring the quality of service is a huge challenge. To solve this problem, this paper proposes an active security defense strategy for container cloud and microservices based on deep reinforcement learning. This method considers both security and quality of service. We model the problem from the dynamic cleaning cycle and the number of microservices replicas, and designed a deep reinforcement learning method to solve it. The results show that the proposed method has better defense effectiveness than the DESOM and SmartSCR defense strategy. The defense efficiency is improved by 34.4% and 12.3% respectively, which verifies the feasibility and effectiveness of the proposed strategy.