Performance Optimization And Application Research Of Traditional Databases Business Migrated To Virtualized Environments

After being many years developed,virtualization technology in calculation,storage,network and management has become increasingly mature and widely used in various fields of digital enterprise.But during the process of virtualization in originations such as government,enterprise and university,personnel seldom focus on performance optimization which resulting performance reduced after virtualized.For the traditional database migration problem this paper designed a migration plan to make sure that business-critical application will not occur service interruption in the process of virtualization.Based on the deep research of the main virtualization technology,this thesis proposes a systematic performance optimization solution to resolve virtualization performance issues,by taking SQL Server database in EXSi environment as an example to improve the efficiency of VMs,such as configuring vNUMA,HT architecture load balance vSMP and other policies to improve the efficiency of physical CPU resources;configuring hypervisor memory page exchange,VM memory allocation and reserve strategy to avoid performance degration caused by memory overcommitted,distributing the read and write load of VMs and configuring vFRC technology to optimize the storage performance,using network optimization methods to resolve performance degradation and communication congestion.This thesis uses the testing tool of the main industry standard benchmark to test the performance of VM before the virtualization and after virtualization,original performance and optimized performance.Through comparing these actual testing data,the feasibility and effectiveness of mentioned research work are verified.The application value of this solution is also shown.