Browse Books

Reviewer: Claudiu Bulaceanu

Today, when designing a functional system is a common matter, emphasis is placed on designing mission-critical systems with enhanced reliability and a high degree of safety. This textbook covers architecture and design of fault-tolerant and high-availability systems, from both the theoretical and the practical points of view. The book is divided into eight parts. The first part comprises four chapters, the first of which is an introduction. The second deals with redundancy techniques for hardware, software, and time. Chapter 3 treats evaluation techniques. Chapter 4 deals with design methodologies. Part 2, comprising seven chapters, discusses the architecture of fault-tolerant computers. The first chapter is an introduction to fault-tolerant architectures. The second chapter treats the categorization of applications and related systems. The third chapter is dedicated to several well-known computer architectures such as IBM and VAX. The fourth chapter briefly analyzes several fault-tolerant implementations, such as the ones issued by AT&T, Tandem, Stratus, and VAXft. The fifth chapter discusses some examples of long-life systems, like the ones used in space for Voyager and Galileo. The sixth chapter highlights the critical systems features and characteristics of systems like the one used for the space shuttle. The seventh chapter is a summary of the part. Part 3, consisting of 13 chapters, displays the principles of fault-tolerant multiprocessor and distributed systems. After an introduction and a review of types of parallel processing, topology of interconnection, and programming models, the next chapters deal with several types of fault-tolerance: static redundancy and dynamic redundancy. Other chapters deal with fault detection, recovery strategies, recovery techniques and schemes (such as rollback and forward recovery), and reconfiguration in multiprocessors. Part 4 deals with several case studies regarding fault-tolerant multiprocessors and distributed systems. This part contains ten chapters that include several well-known implementations, such as the Tandem systems, the Stratus XA/r 300 Systems, the Sequoia systems, the Byzantine resilience model, and the space shuttle system. The six chapters of Part 5 are dedicated to experimental analysis of systems dependability. Statistical techniques, the design phase, and the prototype phase are presented, and the operational phase is discussed at length. Part 6 includes five chapters and covers reliability estimation, including element and system reliability, behavioral decomposition, and examples. Part 7, including seven chapters, focuses on fault tolerance in software. It highlights the motivation for this area of study, how to deal with faulty programs, fault-tolerance in software, reliability models for software, acceptance tests, exception handling, and a case study of an extended distributed recovery block. The final part comprises four chapters and is dedicated to system diagnosis. The different flavors of diagnoses are discussed, including the diagnosis under probabilistic models and under bounded models. This is a classroom textbook intended for advanced undergraduate and graduate students in computer science or a related field of study; IT researchers who want to complete their knowledge of systems reliability; designers of reliable computer systems; and IT project managers. The level and structure of the text, as well as its approach to the fault-tolerant systems field, are typical for the university environment. Each part ends with suitable exercises. The best part of the book is its comprehensive and clear explanation of a field not yet well established and with many things still to uncover. The real-life examples and theoretical mathematical support are additional outstanding features . On the down side, the book would have improved by following at least one example of the design of a fault-tolerant system from beginning to end, passing through all the phases outlined in the book.