GOMAD: A Goal Oriented Approach for Modeling and Analyzing Database Deployment

Database deployment automation involves dealing with the diversity of deployment platforms and the panoply of options to improve the efficiency of query workload. Designers must compare different scenarios and alternatives to find a good physical configuration. This requires hardware experimentation on a target platform or a simulation during the physical design phase. We aim to provide a solution that allows designers to reason up the causality of a given choice at the level of conceptual design and realize the selection of a physical configuration in a transparent way for the designers. In order to tackle this problem, we propose an approach based on goaloriented modeling which helps database designers to automatize and ease the traditional workflow dedicated to choosing an appropriate physical configuration. This latter captures a physical design vocabulary at a higher level of abstraction in order to focus on requirement user and application constraints and not on low level implementation details. Our Goal Model with its generator quickly produces the script related to the database materialization and speeds up the database creation process. We demonstrate the feasibility of our approach presenting a prototype tool.