Models of Embedded Computation
Axel Jantsch · Industrial information technology series · 2005
First, we trace the history of models of computation (MoC) from the early days of computing to the current situation. In the beginning models of computation have been used to study what can be computed in principle. Then, they have been used to address the question how much time and resources are required in principle to solve a given problem. For parallel computers various computational models have been developed to capture the essential features of communication and memory architecture of parallel machines. Then we observe, that the situation becomes really complicated for embedded systems due to real-time requirements, heterogeneous architectures and applications, and real-world interaction. We argue that computational models should be devised for aiding the analysis and design of concrete embedded systems by striking a difficult balance between general and abstract MoCs on one hand and accurate and efficient MoCs on the other hand.