附註:Includes bibliographical references (pages 125-138) and index.
摘要:The Testability of Distributed Real-Time Systems starts by collecting and analyzing all principal problems, as well as their interrelations that one has to keep in mind wh4en testing a distributed real-time system. The book discusses them in some detail from the viewpoints of software engineering, distributed systems principles, and real-time system development. These problems are organization, observability, reproducibility, the host/target approach, environment simulation, and (test) representativity. Based on this framework, the book summarizes and evaluates the current work done in this area before going on to argue that the particular system architecture (hardware plus operating system) has a much greater influence on testing than is the case for `ordinary', non-real-time software. The notions of event-triggered and time-triggered system architectures are introduced, and its is shown that time-triggered systems `automatically' (i.e. by the nature of their system architecture) solve or greatly ease solving of some of the problems introduced earlier, i.e. observability, reproducibility, and (partly) representativity. A test methodology is derived for the time-triggered, distributed real-time system MARS. The book describes in detail how the author has taken advantage of its architecture, and shows how the remaining problems can be solved for this particular system architecture. Some experiments conducted to evaluate this test methodology are reported, including the experience gained from them, leading to a description of a number of prototype support tools. The Testability of Distributed Real-Time Systems can be used by both academic and industrial researchers interested in distributed and/or real-time systems, or in software engineering for such systems. This book can also be used as a text in advanced courses on distributed or real-time systems.
