An optimized VM Allocation Strategy to Make a Secure and Energy-Efficient Cloud Against Co-residence Attack

As cloud computing is gaining increasing popularity, it also raises numerous security challenges. One of such challenges is the co-residence attack, in which attackers take advantage of the resource sharing in multi-tenant cloud to perform diverse attacks against their co-residents on the same physical server. In this paper, we propose to defend against such co-residence attacks in Infrastructure as a Service (IaaS) through a secure, workload-balanced, and energy-efficient Virtual Machine (VM) allocation strategy. Specifically, we model the VM allocation problem as an optimization problem where the goal is to minimize (1) the security risks introduced by the co-residence of VMs from multiple users, (2) the overall power consumption and (3) the unbalanced workload among different physical servers. As this optimization problem is NP-hard, we further apply the Ant Colony Optimization (ACO) algorithm, an evolutionary algorithm inspired by natural ant activities, to identify the optimal allocation strategy. Experiment results demonstrated that the proposed scheme can make the multi-tenant cloud secure and power efficient.