Reconfigurable Multi-formalism Models for Performance Analysis of Distributed Systems

The process of multiformalism modeling involves selecting the most appropriate formalism for individual system components, while ensuring the preservation of overall system coherence. The increasing complexity and adaptability of contemporary systems necessitate the development of dynamic models capable of addressing these challenges effectively. In this context, we propose a framework predicated on Maude for the construction of reconfigurable multiformalism models. Two alternative solutions to this framework are presented. A server management case study, employing stochastic Petri nets and multiclass queuing networks, serves to demonstrate the feasibility of this approach. Empirical experiments indicate that the integration of rewriting techniques within multiformalism modeling has the potential to enhance both the expressiveness and evaluation efficiency of the models.