The Research Of A Self-adaption Performance Guarantee Method Supported By Multi-strategy For SaaS Systems

SaaS (software as a service) is an innovative software application mode, along with the development of Internet technology and the application software. Compared to the traditional web software systems, the SaaS system has many characteristic because of its different application field and the way to realize. Firstly, the SaaS system mainly supports the enterprise key business to enterprise users, which requires the SaaS system should have high reliability in any load environment; Secondly, the SaaS system’characteristics of pay-on-demand and classify-on-tenancy requirement SaaS service providers can provide different levels of service quality to the tenants according to the lease; Finally, most businesses the SaaS system provided have the transaction characteristic, and one request’s failure in the transaction will lead to the whole business failure. Therefore, the characteristics and reliability requirements of SaaS systems call for a set of SaaS systems’performance security mechanism in order to ensure the service quality of high level tenants, improve the processing success rate of business transactions and maximize the gain to cost rate, which continually monitoring system, adjusting and choosing the performance guarantee strategy to meet the business needs of the tenants in the environment of load dynamic changes.Basing on above-mentioned background, firstly this thesis briefly introduces the SaaS system’s architecture and researches the traditional system performance, secondly analyses the targets and strategic characteristics of a variety of performance guarantee strategies and researches the combination mechanism of guarantee strategies. Thirdly, this thesis proposed specific performance analysis methods of system based on resources utilization, and transaction throughput rate of SaaS systems. What’s more, the thesis analysis the basic considerations and process of SaaS system performance guarantee mechanism, and proposed the SaaS system adaptive performance security mechanism, including a series of strategies of system performance evaluation, request access control and request scheduling. In order to improve the accuracy of the SaaS system performance guarantee strategy, the thesis introduces fuzzy control and fuzzy evaluation theory for the core modules of request access and scheduling, researches the request access control algorithm and the request scheduling algorithm based on fuzzy theory.Finally, the article establishes the contrast experiment based on the traditional performance guarantee strategies and the SaaS system performance security strategy, and then analyzes the experiment result based on the average transaction response time and transaction failure rate. The results of experiment proved that the performance guarantee strategies mentioned in this thesis is effective in the SaaS systems.