Quantitative Study Of Availability Assurance And Assessment Of Cloud Computing Based On MRP_CFR

With the expansion of the application scope of cloud computing, the concern about the availability of cloud computing is growing on the mean time. Because of the inherent complexity of availability, there are short of both the effective availability assurance method and the quantitative assessment means. In addition, different users of cloud computing will bring in kinds of unique needs and a set of cloud computing oriented effective availability assurance and quantitative assessment methods are urged.Based on the study of current availability assurance methods, the related factors of cloud computing are sorted out. Aiming to the different failure rules of different resources, the Multi_queuing Resource Provisioning Mechanism Based on Classified Failure Rules is proposed. For the short of quantitative assessment model of availability, an assessment model and the calculation method are given out. The experimental data shows the resource provisioning mechanism and the quantitative assessment method are quite effective.First of all, a study about the current research status of cloud computing is conducted, including the basic concept of cloud computing, its availability, the summary of the characteristics of current availability assurance and assessment methods, especially their weak points. And then, the actual deployment of cloud services in some unit’s inside three layer networks is listed, both of the problems and the advantages which are brought in by its clear and precise network topology structure are discussed.Secondly, aiming to the different resources needed by different levels of cloud services, the MRP_CFR mechanism is proposed, which takes the characteristics of the three layer networks and the cloud services deployed on it into account. MRP_CFR divides the resources needed by cloud computing into computing resources and network interactive resource in general, and analysis the failure rules of each of the resources. The resource allocation is based on the two level multi_queue constructed by different failure rules of different resource types, which supports the priority services to ensure the efficient use of resources and special needs of part of the business.Thirdly, according to the principle of stratified services, we define the concept of SU, SS etc, proposed a quantitative calculation model of cloud computing availability which is based on hierarchical structure and the specific examples of calculation. The summary process is as follows: 1) finish the quantitative calculation of the availability of stratification depending on the services deployment topology; 2) drill accumulate of each stratification availability; 3) give out the overall quantitative availability.At last, according to our resource allocation mechanism MRP_CFR and quantitative availability calculation method, the realization of the key functional modules including event generator, virtual resource pool, the controller and the quantitative calculation of availability is finished based on laboratory environment. And then, the assessment based on real event data and services deployment topology is carried out. Finally, in two application scenarios, the experimental data is acquired and by analyzing the data, the validity of our resource allocation mechanism and quantitative calculation method of availability are verified.