Towards Scalable Validation of Low-Code System Models: Mapping EVL to VIATRA Patterns

Abstract

Adoption of low-code engineering in complex enterprise applications also increases the size of the underlying models. In such cases, the increasing complexity of the applications and the growing size of the underlying artefacts, various scalability challenges might arise for low-code platforms. Task-specific programming languages, such as OCL and EOL, are tailored to manage the underlying models. Existing model management languages have significant performance impact when it comes to complex queries operating over large-scale models reaching magnitudes of millions of elements in size. We propose an approach for automatically mapping expressions in Epsilon validation programs to VIATRA graph patterns to make the validation of large-scale low-code system models scalable by leveraging the incremental execution engine of VIATRA. Finally, we evaluate the performance of the proposed approach on large Java models of the Eclipse source code. Our results show performance speed-up up to 1481x compared to the sequential execution in Epsilon.

Towards Efficient Model Comparison using Automated Program Rewriting

“…Program rewriting has proven to be beneficial for various optimization purposes, as demonstrated in [24] where it was utilized for optimizing type level model queries. Additionally, rewriting has played a crucial role in translating EOL expressions to Viatra for incremental evaluation [6], as well as converting them to MySQL queries for efficient execution on relational databases [7].…”

Selective Traceability for Rule-Based Model-to-Model Transformations

“…Static analysis has been used for enabling the translation from EOL to SQL [9], for optimisation of programs over EMF models [27] and for enabling the translation from EOL to Viatra patterns [8].…”