Towards Virus Scanning as a Service in Mobile Cloud Computing: Energy-Efficient Dispatching Policy Under ${N}$ -Version Protection

Abstract

In this paper, we present a cloud-based architecture of virus scanning as a service (VSaaS) for resource-poor mobile devices. Specifically, we investigate energy-efficient dispatching policy for VSaaS under $N$ -version protection, which dispatches a file to multiple service engines for scanning in parallel to reduce the missed detection error cost. However, the benefit of $N$ -version protection is at the cost of increasing energy consumption and the queue length of service engines. To manage this tradeoff, we aim to reduce the energy consumption of service engines and the missed detection error cost, while satisfying the queue stability for VSaaS. We formulate the file-scan dispatching policy as a stochastic optimization problem, and then propose an online algorithm to dispatch file-scan requests under $N$ -version protection. We show that we can achieve the penalty function, i.e., the combination of the energy consumption and the missed detection error cost, arbitrarily close to the minimum but at the cost of increasing the queue length. Moreover, we compare the proposed dispatching algorithm with the $K$ -least delay algorithm and show by simulation that the proposed dispatching algorithm is more energy efficient to provide $N$ -version protection for VSaaS.