Research And Implementation Of MongoDB Multi Tenant Technology

With the rapid development and increasing maturity of cloud computing technology,SaaS services are more and more widely used in daily production life,and database has also started the road of "going to the cloud".Compared with the local deployment of database,cloud database has the advantages of on-demand use,high availability,high scalability and low cost,the use of single instance and multiple leases can largely reduce the operating costs of SMEs and cloud service providers.Therefore,cloud database has broad prospects.In the process of database "going to the cloud",the first problem to be solved is the sharing and isolation of multi tenant,that is,how to ensure the data isolation between different tenants when multiple tenants share the same database instance.At present,the solution to this problem mainly uses virtualization technology for cloud database deploy-ment,and the current research on multi tenant database is still based on traditional rela-tional database,while the research on multi tenant technology of NoSQL database is rare.Therefore,based on the existing research work of database multi tenant technology,a database system supporting multi tenant is designed and implemented based on mongodb database.The most important feature of multi tenant applications is "single instance,multi ten-ancy",that is,the sharing of applications and the isolation of tenants.This paper imple-ments multi-tenant sharing of database instances based on the "shared database,separate schema" data storage approach in multi-tenant technology.In order to provide data isola-tion when multiple tenants share the same database instance,this paper uses a combination of metadata management and namespace mechanisms to ensure the integrity and security of each tenant's data.At the same time,in order to ensure the normal operation of multi-tenant functions and fast response to tenant requests under concurrent situations,this paper adds a classification management policy to the traditional priority-based scheduling pol-icy to classify tenant requests and then schedule them,which improves the efficient use of resources for multi-tenant management services and improves the availability of the whole system.Finally,the paper tests the final multi-tenant database system,including functional tests,read/write performance tests and concurrency performance tests,and compared with the native MongoDB database.The test results show that the multi-tenant database system implemented in this paper provides basic data access and data isolation functions,and has good performance compared with the original MongoDB database,which can meet some simple multi-tenant database usage scenarios.