Research On Regeneration Method For Distributed Virtual System In Cloud Environment

Distributed virtual system aging under the cloud environment is long time continuous operation on a host computer,lead to the virtual machine and the application status degradation and performance degradation, eventually also lead to the failure of the virtual system. And interruption of service, thus causing heavy losses.In an aging virtual system, there are several aspects:first, according to the regeneration is different, use different regeneration methods also has different effect, so we need to select regeneration methods for the specific circumstances.Second, during the regeneration process, there are correlations between the applications, if regenerate one application may cause other applications running error.Third, how to improve the original migrate VM regeneration, in order to reduce regeneration time and regeneration costs.Fourth,it is difficult to obtain resource usage and consumption in the virtual system, forecasting aging state through resource consumption is very difficult.For the above problem,based on analyse and study traditional software aging treatment, this dissertation include following aspects: first,this dissertation presents regeneration method selection framework based on fuzzy comprehensive evaluation. Using fuzzy comprehensive evaluation When choosing regeneration method, in order to select the most suitable regeneration method.Second, this dissertation presents application restart group based on applications correlation.During the regeneration consider applications correlation, restart the applications in the application restart group, Avoid the application error because of the regeneration.Third, improved regeneration based on migration.Improve traditional Migrate-VM regeneration method, and using the optimization algorithm in the migration process.In order to reduce the regeneration time and costs. At least.design an ageing regeneration prototype system based on response time.Through the response time prediction and detection system aging state, and perform the regeneration.The experimental results show that the proposed improved regeneration method based on migration, has certain feasibility and validity.By comparing the average VM availability, resource utilization, VM migration quantity, normally VM closed number four evaluation parameters, verify the effectiveness of improved regeneration method based on migration.This dissertation start reach from perspective of cloud environment distributed virtual system, and manage aging problem according to the characteristics of cloud environment, and put forward several cloud environment distributed virtual system aging treatment methods.this dissertation makes contribution to reach of software aging problem in cloud environment.