面向高性能计算环境的多维自适应授权访问策略OA北大核心CSTPCD

High-performance computing(HPC)capability is an important manifestation of a country's comprehensive strength and innovation capability,and is one of the key technologies supporting the sustainable development of sci-ence and technology in China.With the development of high-performance computing,more and more researchers in the field have started to pay attention to the HPC environment.Now the HPC environment is facing challenges such as limited resources and increasing number of accounts.In order to ensure the security of the environment and im-prove the utilization of the environment resources,certain authorized access policies need to be set to constrain the access behavior of users.In this paper,a multi-dimensional adaptive access control(MAAC)policy is designed and implemented.The policy is based on machine learning algorithms to analyze user behavior and obtain relevant attributes for users and application communities or business platforms,which are served by the HPC environment.Experimental results show that MAAC can achieve effective and flexible access control to environmental resources.Meanwhile the determination time of the MAAC can be controlled within 1 ms,which is negligible compared with the response time.