The Spring architecture

Abstract

A research approach which challenges several basic assumptions on which most current real-time systems are built is presented. A novel paradigm based on the notion of predictability and on a method for online dynamic guarantee of certain types of deadlines is proposed. The Spring paradigm requires an integrated set of solutions, ranging from design and specification methods and tools, to real-time languages, real-time operating systems, and real-time architectures. An overview of the major ideas of this paradigm is given, and the ramifications of this paradigm for the architecture are explained. In particular, the functional level of the architecture as well as the component level are discussed with respect to the application CPUs, MMUs, floating point coprocessors, DSP chips, other I/O front-end processors, buses, and a specialized scheduling processor. The impact of the Spring paradigm on the development of predictable systems is examined.