ABSTRACT
Real-time distributed computer systems combine virtual environment (VE)– based systems, real-time peer-to-peer (P2P)–based systems, quality-of-service (QoS)–aware distributed computer systems, and hybrid systems (the systems that incorporate both real-time and non-real-time behaviors). Efficient system design for such systems requires a sound framework that can facilitate effective design modeling and simulation (M&S) as well as support for flexible testing of design alternatives. Traditional approaches to system design do not use formal system specification languages or formalisms and are based on accumulated domain knowledge. Thus, system design, in many cases, is separated from system implementation, system testing, and system validation. With the rapid 36advances of distributed real-time computer systems, the complexity of such systems is increasingly challenging traditional nonformalized design approaches. Therefore, formal design methods have been widely used for system design in recent years, and such efforts have been verified by many researchers as suitable for large-scale and complex systems. Among different formal approaches, it is worth noting that model-based system design has proved to be one of the most efficient methods to address the key concerns in complex system design [1–4]. Indeed, the model-based design approach uses a formal language to describe system design models, and such design models are then simulated to predict the performance of the system in real-world scenarios. In particular, the model-based formal design using Discrete Event System Specification (DEVS) [5–7] differentiates it from most other approaches because of its use of a unique integrative framework that can address most of the issues faced in complex system design, and in particular, distributed real-time computer systems.
