A multi-tenant database framework for software and cloud computing applications

University of Technology, Sydney. Faculty of Engineering and Information Technology.Cloud Computing is a new computing paradigm that transforms accessing computing resources from internal data centres to external service providers. This approach is rapidly becoming a standard for offering cost effective and elastic computing services that are used over the internet. Software as a service (SaaS) is one of the Cloud Computing service models that exploits economies of scale for SaaS service providers by offering the same software and computing environment for multiple tenants. This contemporary multi-tenant service requires a multi-tenant database design that can accommodate data for multiple tenants in one single database schema. Due to multi-tenant database resource sharing in this service, the multi-tenant schema should be highly secured, optimized, configurable, and extendable during runtime execution to fulfil the applications’ requirements of different tenants. However, traditional Relational Database Management Systems (RDBMS) do not support such multi-tenant database schema capabilities, and it is a significant challenge to enable RDBMS to support these capabilities. Therefore, one solution is using an intermediate software layer that mediates multi-tenant applications and RDBMS, to convert multi-tenant queries into regular database queries, and to execute them in a RDBMS. Developing such a multi-tenant software layer to manage and access tenants’ data is a hard and complex problem to solve and has significant complexities that involve longer development lifecycle. There are two main contributions of this thesis. Firstly, a proposal for a novel multi-tenant schema technique called Elastic Extension Tables (EET). Secondly, a proposal for a multi-tenant database framework prototype to implement EET schema in a RDBMS. This approach can be used to develop a software layer that mediates software applications and a RDBMS. This software layer aims to facilitate the development of software applications, and multi-tenant SaaS and Big Data applications for both cloud service providers and their tenants. Extensive experiments were conducted to evaluate the feasibility and effectiveness of EET multi-tenant database schema by comparing it with Universal Table Schema Mapping (UTSM), which is commercially used. Significant performance improvements obtained using EET when compared to UTSM, makes the EET schema a good candidate for implementing multi-tenant databases and multi-tenant applications. Furthermore, the prototype of the EET framework was developed, and several experiments were performed to verify the practicability and the effectiveness of using this framework that based on EET multi-tenant database schema. The results of the experiments indicate that the EET framework is suitable for the development of software applications in general, and multi-tenant SaaS and Big Data applications in particular