Concurrent Editing for Multi-Model Environments

Context: Complex systems, such as cyber-physical systems, are developed by multiple developers working on multiple models. To build the actual system, the models need to be consistent to each other. Automatically preserving consistency in such modeling environments introduces further conflicts that occur when developers concurrently edit their models. In addition, enforcing consistency at all times, when it only needs to be guaranteed at build time, may unnecessarily restrict developers. Objective: In our thesis, we plan to develop editing techniques for multi-model environments that provide strict consistency, as well as fast response times under relaxed consistency. Further, we plan to integrate suitable mechanisms for restoring consistency. Method: We adapt the transaction concept from databases, by formalizing consistency-preserving edits as transactions, and implementing transactional editing algorithms. We also incorporate consistency relaxations into the transaction model, and investigate the use of Conflict-Free Replicated Data Types (CRDTs). For restoring consistency, we investigate how to incorporate model repair techniques, as well as using suitable, inconsistency-showing views. Evaluation: We propose to evaluate the performance of the scheduling algorithms through creating a benchmark from existing graph models and edits. Additionally, we plan to evaluate the inconsistency views, and the switch between transactions and CRDTs, with case studies.