面向综合性云平台的主动安全防御策略研究OA
Research on active security defense strategy for comprehensive cloud platform
在容器云与微服务架构环境下,综合性云计算平台面临的攻击手段变得更为复杂且多变.在保障服务质量的同时,如何强化系统防御能力,是当前大型综合计算系统所面临的核心问题.为解决这一问题,提出一种基于深度学习的具有防御成本属性刻画的容器云和微服务主动安全防御策略.通过对动态清理周期与微服务副本数量两个关键维度进行建模,设计一种深度强化学习算法.以中国石油冀东油田云平台为实测算例的实验结果表明,相较于SmartSCR和ASDS-DRL两种防御策略,所提方法在防御效果上具有一定优势,防御效能分别提升了72%和7.5%,验证了所提策略的可行性与有效性.
In container cloud and microservice architecture environments,the attack methods faced by comprehensive cloud computing platforms have become more complex and varied.How to strengthen system defense capabilities while ensuring service quality is currently a major problem faced by large-scale integrated computing systems.In order to solve this problem,a container cloud and microservice active security defense strategy based on deep learning with defense cost attribute characterization is proposed.A deep reinforcement learning algorithm is designed by modeling the two key dimensions of dynamic cleaning cycle and the number of microservice replicas.The experimental results based on the actual measurement example of the cloud platform of Petro China Jidong Oilfield show that,in comparison with two defense strategies SmartSCR and ASDS-DRL,the proposed method has certain advantages in defense effectiveness,with defense efficiency increased by 72%and 7.5%respectively,proving the feasibility and effectiveness of the proposed strategy.
徐震;李波;任毅;苏宏伟;常灿
河北工业大学 计算机科学与软件学院,天津 300401||中国石油冀东油田分公司 勘察设计与信息化研究院,河北 唐山 063004中国石油冀东油田分公司 勘察设计与信息化研究院,河北 唐山 063004中国石油冀东油田分公司 勘察设计与信息化研究院,河北 唐山 063004中国石油冀东油田分公司 勘察设计与信息化研究院,河北 唐山 063004中国石油冀东油田分公司 勘察设计与信息化研究院,河北 唐山 063004
信息技术与安全科学
容器云微服务架构主动安全防御深度强化学习动态清理周期微服务副本数
container cloudmicroservice architectureactive security defensedeep reinforcement learningdynamic cleanup cyclenumber of microservice replicas
《现代电子技术》 2026 (2)
59-64,6
天津市自然科学基金项目(22JCQNJC00160)中国石油冀东油田勘察设计与信息化研究院2023年度建设项目([2023]35号)
评论