4. Process Models

This chapter presents road maps for how to build and execute a distributed simulation in a model-driven way. In this respect, we introduce process models for distributed simulation development and execution. First, we introduce the standards FEDEP and DSEEP and then we present a process model enhanced with a model driven engineering (MDE) approach. Current distributed simulation development and execution process models generally focus on the processes and data that flow among them. They picture the simulation development and execution in terms of activities, information flow, and products that answer what needs to be done by prescribing which workproducts to produce. In general terms, they follow a well-known waterfall software engineering paradigm, where one activity follows another. Although those process models are well suited for guidance for the simulation development and execution, they generally fall short of supporting engineering processes in terms of automation, tool development, and code generation. On the other hand, MDE presents new opportunities for distributed simulation development and execution. First of all, an MDE-based process model sees the models and transformations among them as primary workproducts in the system development lifecycle. So, the focus is on designing and developing models and specifying transformations. Second, well-defined transformations facilitate automation. So, ideally, a model can be transformed into another model, at least semi-automatically with user intervention. Definition of transformation is similar to programming; it requires setting the aim and design for that end. As all artifacts, save for the executable codes and supporting files, are models, they can be reused in many ways. For instance, a federation architecture model can be transformed into source code in a preferred programming language, where the transformation is the code generation. Reusing the same model, we can generate code in another language by crafting another transformation.