# McFSM: Globally Taming Complex Systems

## TL;DR

The paper introduces McFSMs, a novel modeling approach for complex, interdependent cyber-physical systems, providing a global description, theoretical foundation, and tools for efficient implementation and correctness verification.

## Contribution

It presents McFSMs, a new method with a domain-specific language and model-driven tools to model, analyze, and generate code for complex interdependent systems.

## Key findings

- Provides a formal foundation for McFSMs.
- Enables automatic code generation in multiple languages.
- Facilitates correctness verification at compile-time.

## Abstract

Industrial computing devices, in particular cyber-physical, real-time and safety-critical systems, focus on reacting to external events and the need to cooperate with other devices to create a functional system. They are often implemented with languages that focus on a simple, local description of how a component reacts to external input data and stimuli. Despite the trend in modern software architectures to structure systems into largely independent components, the remaining interdependencies still create rich behavioural dynamics even for small systems. Standard and industrial programming approaches do usually not model or extensively describe the global properties of an entire system. Although a large number of approaches to solve this dilemma have been suggested, it remains a hard and error-prone task to implement systems with complex interdependencies correctly. We introduce multiple coupled finite state machines (McFSMs), a novel mechanism that allows us to model and manage such interdependencies. It is based on a consistent, well-structured and simple global description. A sound theoretical foundation is provided, and associated tools allow us to generate efficient low-level code in various programming languages using model-driven techniques. We also present a domain specific language to express McFSMs and their connections to other systems, to model their dynamic behaviour, and to investigate their efficiency and correctness at compile-time.