An Adaptive VM Allocation Approach for Mitigating Co-resident Attack in Cloud Based on Improved NSGA-II

Xin Yang, Jintao Chen, Kuiye Ding, Chao Liu, Jiashen Liu, Songyao Hou

Abstract: Cloud computing allows users to execute tasks on virtual machines (VMs) providing flexible and cost-effective services, but bring new risks due to the shared resources among multiple tenants. Malicious tenants have the potential to build side channels to steal private information from legitimate users while co-residing with them. VM migration technique defends against co-resident attack by dynamically migrating takes among multiple servers. However, VM migration protect tasks at the cost of degrading its performance or energy consumption. The inefficient VM migration leads to excessive power consumption, resource wastage, even security breaches. In this paper, we described the executing and migrating process of as a bridge crossing model. On this basis, we propose an adaptive VM migration and allocation architecture centered in task execution and based on user threat perception. We improve NSGA-II algorithm to optimize the migrating and allocating process, incorporating with cluster-based user threats model to capture the fine-grained user information details and assess the threats of users. Our multi-dimension assessment evaluates the defensive capacity and effectiveness for different scenarios. The simulation results show that the proposed method is efficient in mitigating co-resident attack.