Research On Evaluation And Selection Of Cloud Service Under The Specific Requirements

In recent years, the development and application of the Internet technology was very quickly and very widely. This has promoted the emergence and development of cloud computing technology. Under the leadership of the world of large Internet Co, which not only made cloud computing technology developing rapidly, but also made cloud computing services increasing quickly. Then one of the key problems in current research of cloud computing is that the users how to choose the high quality services according to their own needs.In this paper, through the analysis of cloud service status evaluation and selection of service, many studies showed that weight of each attribute value and attribute of the service of the actual value is irrelevant. So most of the evaluation models with fixed weights, which not only will lead to the selected services don’t meet the actual needs but also will lead to service failures or SLA the default. The no improvement classical algorithms are can’t guarantee the quality and efficiency of the cloud service portfolio selection. In order to solve the above problems, the key problem of the following research in this article is the cloud service QoS evaluation and service portfolio selection.Firstly, this paper analyzes current QoS evaluation model. The defects of fixed weight in the evaluation are pointed out. An evaluation method based on the theory of variable weight is proposed. On the basis of the theory of variable weight, QoS variable weight vector that according to the characteristics of the current field of users is established. After the attribute weights of each service are adjusted through variable weight vector. Accuracy of the comprehensive evaluation is improved.Secondly, the deficiency of the traditional particle swarm algorithm are analyzed and summed up. In order to improve the convergence speed of particle swarm algorithm that the way of adjusting value of the inertia weight and learning factor are used. Through the related experiments of the improved algorithm parameter values are determined so that the improved algorithm optimization effect is better. The operation can make the particle velocity and position update faster. So we can choose a better combination of cloud services.Lastly, cloud services portfolio selection model which includes the constraint condition and objective function is given in this paper according to the characteristics and needs of the user’s field. The simulation experiment is about that selection algorithm for service composition before and after the improvement of the efficiency of selection and the way of the evaluation before and after the improvement of QoS services composition. Through the comparison of experimental result analysis found that two kinds of method in the improved effect is more obvious.