CET: Clustering Extension Table research in multi-tenant database for SaaS applications

Abstract

Microsoft's ET (extension table) approach makes the data model arbitrarily extensible while retaining the cost benefits of using a shared database. But it adds complexity for database querying and updating. We improve it and propose CET (Clustering Extension Table). There are several tables in CET, including BT (Basic Table), CT (Clustering Table), CMT (Metadata Table for CT), ET (Extension Table) and EMT (Metadata Table for ET). The similarity of tenants is adopted to cluster them into different groups. In the group, some data in the basic table and extension table is transformed and consolidated together into the CT (Clustering Table). In one group, CT extracts the common similarity of the tenants in the group to avoid the complexity of using ET. Then SER (Schema Efficiency Ratio) and DER (Data Efficiency Ratio) are described to control tenants' adscription into their groups. Experiments are also conducted to show the improvement of ET.