Schema-Mapping Techniques On Multi-Tenant Databases For Software As A Service

Schema-Mapping Techniques on Multi-Tenant Databases for Software as a ServiceAlong with the rapid development of Internet technology and the application, the rising demand of the software market and the rising productivity of software, software transfers from the traditional products into services. SaaS(Software as a Service) really captured this idea, and thoroughly achieved the objectives of transferring the product into service. In the development process of SaaS. due to hardware、software, networking, system maintenance and personnel costs, resulting in that multi-tenant software prompted much concern, because all tenants share the costs of software、hardware, the development of application and the maintenance of application in the multi-tenant software, which thereby reduce the cost of each tenant. Then in order to meet this high efficiency of the development costs, data storage model and query conversion mechanism in multi-tenant environment becomes particularly important.In this paper, we compare the basic table layout, the private table layout, the extension table layout, the universal table layout, the pivot table layout and the chunk table layout, also combine with the thinking of the Chunk Folding, so we design the algorithm, and give formal description and instructions. According to establishment of the proposed algorithm to design the schema mapping mechanism, mapping most heavily-utilized parts of logical schemas into traditional table and the remaining parts into Chunk Tables to match their structure as closely as possible, and we get good performance. The schema mapping mechanism implement the query conversion between the logical data schema and the physical storage schema in tenants, this conversion takes place in the data storage mode of a shared database shared data schema. To illustrate the effectiveness of Chunk Folding, the paper describes the experimental platform for multi-tenant database, at the same time; we make the corresponding tests for query translation.Finally, the experimental platform of database has been built, it include extension table and base table frame to simulate hosting CRM service. The algorithm of Chunk Folding has been implemented by thinking about logic schema of tenants, data of allocation and query of application. Experiments show that access efficiency of data center of multi tenants is better than traditional relational database.