An optimal resources scheduling strategy on multimedia cloud computing under multi- device constraint

In view of the problem of inaccurate scheduling by using traditional resource scheduling method, because the method is mainly based on extracting and classifying the resource features to make scheduling, ignoring the effect of the feature relevance between the resources on the scheduling results. This paper presents a model for multimedia cloud resource scheduling based on multi- device constraint. In this method the objective function is no longer constrained only by the CPU computing capacity and the minimized completion time, but to achieve a minimum time-consuming of CPU, memory and other peripherals operation are considered as the scheduling objectives. Then the utilization of solving constrained jointly is employed to obtain the mapping relationship of the optimal virtual and physical machine. Moreover, a regressive dimensionality reduction algorithm is designed for this scheduling model to solve the high dimensional problems aroused by multi-device constraints. Simulation results show that the improved algorithm has a better performance than the traditional algorithm, has a good efficiency and has a certain convergence.