Research And Application Of Cloud Service Evaluation Based On Improved TOPSIS

In recent years,cloud computing technology has been the key technology for providing IT services,and aims to provide large-scale and flexible IT resource services through the Internet.After the rapid development and widespread application of cloud computing technology,many cloud service providers are competing to develop a large number of similarly functional service products for cloud users to choose.For cloud users,how to scientifically and reasonably evaluate the advantages and disadvantages of these cloud services has been becoming a very challenging task.Therefore,we must find a way to distinguish the advantages and disadvantages of cloud services.Service quality evaluation is a traditional service evaluation method.Quantitative evaluation of cloud services can be effectively carried out by establishing a suitable QoS model for cloud features.This paper focuses on cloud service evaluation,the screening of cloud service QoS evaluation criteria,the establishment of cloud service QoS model,the improvement of TOPSIS evaluation methods and the application of in cloud service evaluation of improved TOPSIS.The main work of this paper is as follows:(1)For the cloud service evaluation under different application scenarios,a dynamic screening method for evaluation criteria is proposed and basing on dynamic screening evaluation criteria,Cloud Service Quality Model is established —CSERVQUAL Model.Fully combine the service characteristics and service processes in different cloud application scenarios,and establish a cloud service QoS evaluation index system.Combine fuzzy prioritization method with the maximal irrelevant method to screen the evaluation criteria according to the quality dimension priority and the degree of correlation between the indicators in different cloud service application scenarios,in order to screen out important Evaluation indicators with high importance and low correlation.Based on the dynamic screening of evaluation indicators,a four-dimensional cloud service QoS model with cloud service security,service performance,service characteristics and commerciality as the quality dimension is established.The measurement indicators selected under each quality dimension are based on cloud service application scenarios and user preference dynamics.The changes and experiments verify that the CSERVQUAL model established in different cloud service scenarios is more suitable in cloud service evaluation.(2)As to the reverse order problem existing in the traditional TOPSIS method,we propose a improved TOPSIS based on improving efficacy coefficient —IEC-TOPSIS algorithm.Firstly,the reasons for the reverse order phenomenon are analyzed;then,the efficacy coefficient is improved by respectively using the historical best and worst data of the index as the satisfaction value and the impermissible value.To encourage the cloud service provider to provide high quality of cloud service,the service provider incentive factor is added.The improved efficiency coefficient performs dimensionless processing on the criteria,and replaces the positive and negative ideal solutions with absolute positive and negative ideal solutions to eliminate the reverse order phenomenon in the traditional TOPSIS method.The effectiveness,isotonicity and optimization of algorithm performance in the application of cloud service evaluation are verified by real service quality dataset.(3)In order to verify the adaptability of the CSERVQUAL model and the effectiveness of the IEC-TOPSIS evaluation method,this paper applies them to the cloud service evaluation system and implements the system.The results show that in different cloud service application scenarios,the application of CSERVQUAL model and IEC-TOPSIS evaluation method we proposed in cloud service evaluation can provide effective decision support and more targeted for cloud users.