Research On The Method Of Dynamic Adaptive Resource Adjustment On The Cloud Services SLA Performance Constraint

With the development of cloud computing,facing component serve of cloud application system has produced.The cloud application system provides services to users in the form of business flow,different tasks are assigned to different component in the different forms of business flow.The application system is established by many component servers.When system performance declines,resource adjustment is needed to improve system performance for SLA constraint.The method of traditional resource adjustment faces VMs,not thinking about resource need of component,lacking SLA constraint based on component.Those problems will lead to resource adjustment inaccurate and waste resource.How to adjust resource becomes a serious problem in the cloud computing.To solve the problems caused by the traditional method of resource adjustment,this thesis proposes a method of dynamic adaptive resource adjustment on the cloud services SLA performance constraint.Resource dynamic adjustment includes in calculating component goal response time for SLA constraint,selecting key component,selecting resource type and calculating resource number.Firstly,analysis the historical data of component response time and application response time,and build the relation model between the component response time and the SLA constraint and the Bayes classifier for the application response time transforming the component response time of SLA constraint.Secondly,build the selecting model based on component goal response time,component invoking rate and user concurrency,and select key component for forming key component set based on optimization algorithm.Thirdly,analysis the historical data of component response time and four resource types using number,and processing data,calculate first difference,calculate correlation coefficient,calculate relevancy.Then select adjusted resource type based on gray relation analysis.Fourthly,build the relation model between component response time and resource using number,and confirm the equation between component response time and resource using number,calculate resource adjustment number based on component response time for SLA constraint.Then many resource adjustment project can be produced,and calculate cost benefit of ever resource adjustment project,select best resource adjustment project for executing resource dynamic adjustment.Finally,use the experiment to proof the effectiveness,the feasibility and the superiority of the method proposed in this thesis.Experimental results show that the method proposed in this thesis can effectively calculate component goal response time for SLA constraint,and select key component and adjusted resource type.At the same time calculate resource adjustment number and select best resource adjustment project using comparison experiment to proof the effectiveness,the feasibility and the superiority of the method proposed in this thesis.