A system is defined to be dependable if reliance can be justifiably placed on the service it delivers. It is a definition adopted a few decades ago, originating from the work of the IFIP Working Group 10.4 on Dependable Computing and Fault Tolerance. It has been widely recognized since then, and used in most major works. The concept of dependability involves several attributes, which make the practical use of dependability easier. They primarily include reliability, safety, and security.

This book is a collection of papers selected from two annual workshops organized in 2008: the Workshop on Architecting Dependable Systems (WADS) and the Workshop on Views on Designing Complex Architectures (VODCA). The papers are organized in three parts, covering service-oriented architectures (SOA), fault tolerance and assessment, and security-related issues.

The part on SOA gives a few hints on how to proceed to ensure dependability for a large system, where the number of variables and their mutual relationships is too big to successfully tackle the problems theoretically. The authors of the various papers make suggestions to use architectural patterns, architectural adaptation to environmental changes, controlled relinquishing of design constraints, and benchmarking.

While SOA is a relatively new area of research, which essentially came to existence with the advent of the Web, fault tolerance can be considered as old as computing itself. The new aspect of fault tolerance considered in this section of papers is addressing dependability attributes of networked systems from the architectural standpoint. Specific topics discussed in this volume include: self-repair without global shutdown, using redundancy to ensure fault tolerance, applying security analysis to verify and improve fault tolerance, and new techniques to deal with architectural complexity.

The third part addresses security issues at the architectural level. This is probably the most compelling task, since one has to deal with multiple layers of security in every computer system. Several theoretical tools are discussed--process algebras, the Paradigm language, subjective logic (SL), and communicating sequential processes (CSP)--with some level of experimentation to make the results more meaningful.

Overall, books that are collections of papers are hard to review, unless the papers are subject overviews discussing the state of the art in a certain discipline, which is not the case with this book. Nevertheless, reading this book has helped me refresh my knowledge of several theoretical concepts related to dependability and find their new applications. I am looking forward to seeing the publication of the next volume in this series.