[share_ebook] Designing And Building Robust, Dependable Real Time Systems

This volume contains the lectures given in honor to Georg Färber as tribute to his contributions in the area of real-time and embedded systems. The chapters of many leading scientists cover a wide range of aspects, like robot or automotive vision systems or medical aspects.

Specification and design methodology has seen significant growth as a research area over the last decade, tracking but lagging behind VLSI design technology in general and the CAD industry in particular. The commercial rush to market tries to leverage existing technology which fuels CAD design tool development. Paralleling this is very active basic and applied research to investigate and move forward rational and effective methodologies for accomplishing digital design, especially in the field of hardware/software codesign.

This book addresses the challenges associated with efficient Mixed-Criticality (MC) system design. We focus on application analysis through execution time analysis and task scheduling analysis in order to execute more low-criticality tasks in the system, i.e., improving the Quality-of-Service (QoS), while guaranteeing the correct execution of high-criticality tasks. Further, this book addresses the challenge of enhancing QoS using parallelism in multi-processor hardware platforms.

Embedded systems are ubiquitous. They appear in cell phones, microwave ovens, refrigerators, consumer electronics, cars, and jets. Some of these embedded s- tems are safety- or security-critical such as in medical equipment, nuclear plants, and X-by-wire control systems in naval, ground and aerospace transportation - hicles. With the continuing shift from hardware to software, embedded systems are increasingly dominated by embedded software. Embedded software is complex. Its engineering inherently involves a mul- disciplinary interplay with the physics of the embedding system or environment. Embedded software also comes in ever larger quantity and diversity. The next generation of premium automobiles will carry around one gigabyte of binary code. The proposed US DDX submarine is e?ectively a ?oating embedded so- ware system, comprising 30 billion lines of code written in over 100 programming languages. Embedded software is expensive. Cost estimates are quoted at around US$15– 30 per line (from commencement to shipping). In the defense realm, costs can range up to $100, while for highly critical applications, such as the Space Shuttle, the cost per line approximates $1,000. In view of the exponential increase in complexity, the projected costs of future embedded software are staggering.