Multiple Sparse Tables based on pivot table for Multi-tenant Data Storage in SaaS

In order to excellently support SaaS application, multi-tenant database system needs to meet the tenants' requirement of isolation and on-demand customization, and then needs to provide data storage mechanism and index mechanism that supporting isolation and flexibility. Multiple Sparse Tables is a good Approach for Multi-tenant Data Storage in SaaS, but no use of physical index provided by RDBMS. Based on the Multiple Sparse Tables approach, in this paper, we proposed a meta-data driven indexing mechanism. According to tenants' customization requirement, the model constructs respective index metadata for business data of the tenants, and achieves isolation of index data & customization; meanwhile the index maintenance strategies are given. According to tenants' access requests and the Pivot Table, the model returns the tenants' result sets more quickly or updates index data on-demand. Detailed experimental results show that the index maintenance and data access of this approach works with good performance under balanced conditions.