Virtual Machines Scheduling Algorithm Oriented Load Forecast

A virtual machines scheduling algorithm oriented load forecast is presented in this paper to improve Credit algorithm, which solves the problem of the adjustment of time slices in virtual machines based on their loads in systems to satisfy the requirements of timeliness, accuracy and fairness. On the basis of shared CPU resources, we propose a progressive time slices adjustment methodology in virtual machines, and based on it, a virtual machines scheduling model based on the minimum deviation of satisfaction degree of resources requirements is designed to adapt to the change of the load forecast results, which has implemented the management of shared CPU resources with a higher efficiency. The results of experiments show that criteria F₁ and smaller fu would enable the executive cycles of virtual machines approach to load forecast.