An Efficient Virtual Machine Management Algorithm for Vehicular Clouds

Abstract

Vehicular cloud computing (VCC) is one of the prominent fields of research, which combines both vehicular ad hoc network and cloud computing. In VCC, the vehicles are acting like virtual machine hosts to process the tasks that are given by the roadside units (RSUs). The vehicles and RSUs coordinate each other to share several information. Further, the RSUs are connected to the cloud through the Internet. As the vehicles under consideration are not residing in the same area or street or zone or grid, the problem of managing virtual machine hosts is very much challenging to process the tasks. In this paper, we introduce virtual machine management algorithm for vehicular clouds. Moreover, we present three variants of the proposed algorithm, namely first fit strategy, best fit strategy and worst fit strategy, respectively. The performance of these variants is evaluated by means of three performance metrics through simulation runs with a wide variety of vehicles, vehicle speeds, length of the grids and data sizes. The result of these variants is extensively compared. The comparison results show the superiority of the best fit strategy among all the variants of the proposed algorithm.