Heterogeneous Model Query Optimisation

With the growing size and complexity of software systems, the underlying models also grow in size proportionally. These large-scale models pose scalability issues for model-driven engineering technologies. These models can be persisted in various backend technologies (such as file systems, document and relational databases) and can be represented in different formats such as XMI and Flexmi. Several tailored high-level model management languages such as OCL and EOL enable developers to work on different backend technologies in a uniform way by shielding them from the complexities of different backends. On the contrary, performance with respect to execution time in tailored model management languages programs becomes one of the major scalability bottlenecks. In this work, we propose an architecture built on top of existing model query languages to facilitate query optimisation. The proposed approach will benefit from compile-time static analysis and automatic program rewriting to optimise queries operating over heterogeneous backend technologies. Optimisation strategies and performance will vary depending on the type of queries and the backend modelling technology. We expect to significantly improve performance (decrease in one order of magnitude of execution time) for model management programs, particularly over large-scale models.