Usability And Reliability Testing-Evaluation And Optimization Of Cloud Service

Regarding the outstanding advantages of on-demand use,metering and rapid elasticity,cloud services have been rapidly popularized,the application fields are becoming more and more extensive,and attract attention from academia and industry.Cloud services are provided through cloud platforms and networks,the usability and reliability directly affect the quality of cloud services,so they are the most important concerns of users and it is of great significance to conduct testing-evaluation and optimization research.Currently,the usability and reliability testing-evaluation method of cloud services are mainly focused on processes monitoring,fault early warning and resources optimization.However,the systematic testing-evaluation and optimization are still in the exploratory stage.The subjectivity and inconsistency of the definition of cloud service quality indexes lead to blindness of testing-evaluation,and it is necessary to unify cloud service quality model to support comprehensive testing-evaluation.The differentiation between the indexes and methods of cloud service usability and reliability testing-evaluation makes its difficult to measure the SLA achievability,and it is important to research the systematic testing-evaluation methods,respectively.Due to the limited edge resources,the cloud services which convergence edge computing are liable to cause service jitter and making the falling of the reliability,and it is need to study the cloud-edge load cooperative migration to ensure the reliability.Therefore,this paper takes the usability and reliability of cloud services as the research object,aims at the systematic testing-evaluation and optimization,and carries out the key technology research from the aspects of cloud service quality model,usability and reliability testing-evaluation methods and edge resource optimization.The main research contents and innovations are as follows:(1)Regarding the inconsistent definitions of cloud service quality indexes which leads to the blindness of evaluation and poor re-usability,this paper analyzing the quality characteristics of cloud services and their internal connections,and establishing and standardizing the cloud service quality model.Based on this model,the cloud services quality assessment function and testing-evaluation method are constructed,and a comprehensive testing-evaluation method which base on analytic hierarchy process is proposed.Finally,it is compared with ITSQM and SMI in detail,and the proposed quality model in this paper takes into account the advantages of SMI's objective testing-evaluation,and expanding the usability of ITSQM,and effectively improves SMI's defect of paying too much attention to cloud computing system and weakening Quality of Experience.(2)Regarding the uncertainty of cloud service usability requirements and the subjectivity of index definition,and this problem results in a limited application scope for cloud services selection and user experience can't be guaranteed effectively,this paper analyzes the behavioral factors which affecting users'choice of cloud services in detail,and proposes the cloud service usability testing-evaluation indexes from functional suitability,specialization,humanization,trust and value.By using TOPSIS and analytic hierarchy process(AHP),a cloud service usability testing-evaluation method is constructed,which supports the automatic generation of differentiated index weights and avoids the influence of subjective malicious factors.The experimental results show that the accuracy of this evaluation method is higher than AHP.(3)Regarding the difference of testing-evaluation indexes and methods which due to the reliability of cloud services is difficult to be effectively guaranteed,the emphasis of this paper is placed on the comprehensive factors which affecting the reliability of cloud services,and constructing the testing-evaluation indexes from availability,continuity,recoverability,and data backup.In order to ensure the consistency of testing-evaluation results through dimensionless and normalized processing,a cloud service reliability testing-evaluation method based on classification statistics and hierarchical variable weight is proposed to eliminate the impact of the difference of evaluation indicators on the comprehensive evaluation results.Finally,an example is given to verify this testing-evaluation method,and the result is the proposed testing-evaluation method has strong applicability and operability,and can provide effective guidance for cloud service reliability guarantee.(4)Regarding the problem of low efficiency of dynamic adjustment of edge resources on demand in cloud-edge collaboration scenario which leads to poor business stability and even interruption,the resource scheduling algorithm for cloud services based on Markov prediction model is proposed by analyzing of cloud-edge collaboration network characteristics,and this algorithm can be used to ensure the reliability of cloud services when the server to be migrated and the transfer link can be quickly selected under the condition of server overload threshold.Finally,the simulation results show that high-load servers can successfully transfer load to low-load servers,and improving the success rate of dynamic allocation of edge resources on demand,ensuring the accessibility rate of cloud services and reducing the failure rate,so as to achieve the goal of ensuring the reliability of cloud services.