Efficient Model Loading through Static Analysis

Growing model sizes reveal scalability issues in the current generation of model management tools and technologies. Hence, there is a demand for more sophisticated mechanisms for execution engines of model management languages to make the most efficient use of the available system resources. In this paper, we present an approach to enable execution engines of model management programs to load models partially and just keep the necessary parts of the model in memory for as long as they are needed. This approach leverages sophisticated static analysis of model management programs to load only relevant model elements instead of naively loading the entire models into memory. In this way, the proposed approach aspires to enable model management programs to process models faster with a reduced memory footprint compared to naive complete model loading which is the current state of the art.