A Flexible Data Model for Multi-tenant Databases for Software as a Service

We study data models for multi-tenant databases for Software as a service (SaaS), which is important for emerging big data applications. In our study we consider the shared table architecture model of SaaS for its advantages over the shared machine and shared process models. Our goal is to further improve the processing time and space utilization of shared table model. For this, we introduce a new data model, called Index table layout and compare its performance with the Chunk table layout. Our model includes an additional ´index´ column in the base table that cross-references via the index across supporting tables. Not needing nulls, the proposed model enjoys quick access of the values in a column. We evaluated the performance of our model, for which we crated randomly large data for Chunk table layout and from which we then constructed the tables for the Index table layout. We also compiled 3 types of queries used in our evaluation. Our results show that Index table layout is better than the Chunk table layout in terms of expressive power, space utilization, and query processing time.