Model-Based Design of Real-Time Systems

Model-based techniques have found widespread use in the design of embedded systems. Models and methods are typically based on well-defined models of computation and allow for building composite components out of simpler components. The advantages of using high-level system abstractions that comprise software as well as the underlying hardware platform lead to an increasing use of the associated design methods: Methods and tools for correctness in terms of qualitative and quantitative requirements, as well as design environments for specification, design space exploration, optimization and code generation. Model-based design is getting increasingly important in the context of parallel and distributed embedded systems due to the tremendous complexity induced by concurrent execution. For real-time systems, it is important to guarantee throughput and latency for each application running on the designed system. Analysis and design techniques that are able to provide such guarantees have been proposed for some of the celebrated platform models and models of computation. In other cases the classical hard real-time scheduling models and algorithms are directly applicable. In the presentation we will classify and link model-based design techniques, models of computation, as well as real-time scheduling theory and algorithms.