Modeling And Analysis Of Dependable Systems

The monographic volume addresses, in a systematic and comprehensive way, the state-of-the-art dependability (reliability, availability, risk and safety, security) of systems, using the Artificial Intelligence framework of Probabilistic Graphical Models (PGM). After a survey about the main concepts and methodologies adopted in dependability analysis, the book discusses the main features of PGM formalisms (like Bayesian and Decision Networks) and the advantages, both in terms of modeling and analysis, with respect to classical formalisms and model languages. Methodologies for deriving PGMs from standard dependability formalisms will be introduced, by pointing out tools able to support such a process. Several case studies will be presented and analyzed to support the suitability of the use of PGMs in the study of dependable systems. Contents:Dependability and ReliabilityProbabilistic Graphical ModelsFrom Fault Trees to Bayesian NetworksFrom Dynamic Fault Tree to Dynamic Bayesian NetworksDecision Theoretic DependabilityThe RADyBaN Tool: Supporting DependabilityCase Study 1: Cascading FailuresCase Study 2: Autonomous Fault Detection, Identification and RecoveryCase Study 3: Security Assessment in Critical InfrastructuresCase Study 4: Dynamic Reliability Keywords:Dependability;Reliability;Probabilistic Graphical Models;Bayesian Networks;Fault Detection Identification and Recovery

Foundations of Dependable Computing: Models and Frameworks for Dependable Systems presents two comprehensive frameworks for reasoning about system dependability, thereby establishing a context for understanding the roles played by specific approaches presented in this book's two companion volumes. It then explores the range of models and analysis methods necessary to design, validate and analyze dependable systems. A companion to this book (published by Kluwer), subtitled Paradigms for Dependable Applications, presents a variety of specific approaches to achieving dependability at the application level. Driven by the higher level fault models of Models and Frameworks for Dependable Systems, and built on the lower level abstractions implemented in a third companion book subtitled System Implementation, these approaches demonstrate how dependability may be tuned to the requirements of an application, the fault environment, and the characteristics of the target platform. Three classes of paradigms are considered: protocol-based paradigms for distributed applications, algorithm-based paradigms for parallel applications, and approaches to exploiting application semantics in embedded real-time control systems. Another companion book (published by Kluwer) subtitled System Implementation, explores the system infrastructure needed to support the various paradigms of Paradigms for Dependable Applications. Approaches to implementing support mechanisms and to incorporating additional appropriate levels of fault detection and fault tolerance at the processor, network, and operating system level are presented. A primary concern at these levels is balancing cost and performance against coverage and overall dependability. As these chapters demonstrate, low overhead, practical solutions are attainable and not necessarily incompatible with performance considerations. The section on innovative compiler support, in particular, demonstrates how the benefits of application specificity may be obtained while reducing hardware cost and run-time overhead.

Over the last two decades, a major challenge for researchers working on modeling and evaluation of computer-based systems has been the assessment of system Non Functional Properties (NFP) such as performance, scalability, dependability and security. In this book, the authors present cutting-edge model-driven techniques for modeling and analysis of software dependability. Most of them are based on the use of UML as software specification language. From the software system specification point of view, such techniques exploit the standard extension mechanisms of UML (i.e., UML profiling). UML profiles enable software engineers to add non-functional properties to the software model, in addition to the functional ones. The authors detail the state of the art on UML profile proposals for dependability specification and rigorously describe the trade-off they accomplish. The focus is mainly on RAMS (reliability, availability, maintainability and safety) properties. Among the existing profiles, they emphasize the DAM (Dependability Analysis and Modeling) profile, which attempts to unify, under a common umbrella, the previous UML profiles from literature, providing capabilities for dependability specification and analysis. In addition, they describe two prominent model-to-model transformation techniques, which support the generation of the analysis model and allow for further assessment of different RAMS properties. Case studies from different domains are also presented, in order to provide practitioners with examples of how to apply the aforementioned techniques. Researchers and students will learn basic dependability concepts and how to model them using UML and its extensions. They will also gain insights into dependability analysis techniques through the use of appropriate modeling formalisms as well as of model-to-model transformation techniques for deriving dependability analysis models from UML specifications. Moreover, software practitioners will find a unified framework for the specification of dependability requirements and properties of UML, and will benefit from the detailed case studies.

This book is intended to provide the interested reader with basic information on various issues of the dependability analysis and evaluation of engineering systems with the principal goal to help the reader perform such an analysis and evaluation. By the definition of the IEC International Standard 50(191) dependability is the collective term used to describe the availability peiformance and its influencing factors: reliability peiformance, maintainability peiformance and maintenance support performance. Dependability is a term used for a general description of system performance but not a quality which could be expressed by a single quantitative measure. There are several other quantitative terms, such as reliability, unreliability, time-specific and steady-state availability and unavailability, which together form a basis for evaluating the dependability of a system. A system is taken as dependable if it satisfies all requirements of the customers with regard to various dependability performances and indices. The dependability deals with failures, repairs, preventive maintenance as well as with costs associated with investment and service interruptions or mission failures. Therefore, it is a very important attribute of system quality. The dependability evaluation is strongly based upon experience and statistical data on the behavior of a system and of its elements. Using past experience with the same or similar systems and elements, the prospective operation may be predicted and improved designs and constructions can be conceived. Hence, the dependability analysis makes it possible to learn from the past for better future solutions.

As software systems become ubiquitous, the issues of dependability become more and more crucial. This state-of-the-art survey contains 18 expanded and peer-reviewed papers based on the carefully selected contributions to the Workshop on Architecting Dependable Systems (WADS 2006) organized at the 2006 International Conference on Dependable Systems and Networks (DSN 2006), held in Philadelphia, PA, USA, in June 2006.

Over the last two decades, a major challenge for researchers working on modeling and evaluation of computer-based systems has been the assessment of system Non Functional Properties (NFP) such as performance, scalability, dependability and security. In this book, the authors present cutting-edge model-driven techniques for modeling and analysis of software dependability. Most of them are based on the use of UML as software specification language. From the software system specification point of view, such techniques exploit the standard extension mechanisms of UML (i.e., UML profiling). UML profiles enable software engineers to add non-functional properties to the software model, in addition to the functional ones. The authors detail the state of the art on UML profile proposals for dependability specification and rigorously describe the trade-off they accomplish. The focus is mainly on RAMS (reliability, availability, maintainability and safety) properties. Among the existing profiles, they emphasize the DAM (Dependability Analysis and Modeling) profile, which attempts to unify, under a common umbrella, the previous UML profiles from literature, providing capabilities for dependability specification and analysis. In addition, they describe two prominent model-to-model transformation techniques, which support the generation of the analysis model and allow for further assessment of different RAMS properties. Case studies from different domains are also presented, in order to provide practitioners with examples of how to apply the aforementioned techniques. Researchers and students will learn basic dependability concepts and how to model them using UML and its extensions. They will also gain insights into dependability analysis techniques through the use of appropriate modeling formalisms as well as of model-to-model transformation techniques for deriving dependability analysis models from UML specifications. Moreover, software practitioners will find a unified framework for the specification of dependability requirements and properties of UML, and will benefit from the detailed case studies.

Foundations of Dependable Computing: Paradigms for Dependable Applications, presents a variety of specific approaches to achieving dependability at the application level. Driven by the higher level fault models of Models and Frameworks for Dependable Systems, and built on the lower level abstractions implemented in a third companion book subtitled System Implementation, these approaches demonstrate how dependability may be tuned to the requirements of an application, the fault environment, and the characteristics of the target platform. Three classes of paradigms are considered: protocol-based paradigms for distributed applications, algorithm-based paradigms for parallel applications, and approaches to exploiting application semantics in embedded real-time control systems. The companion volume subtitled Models and Frameworks for Dependable Systems presents two comprehensive frameworks for reasoning about system dependability, thereby establishing a context for understanding the roles played by specific approaches presented in this book's two companion volumes. It then explores the range of models and analysis methods necessary to design, validate and analyze dependable systems. Another companion book (published by Kluwer) subtitled System Implementation, explores the system infrastructure needed to support the various paradigms of Paradigms for Dependable Applications. Approaches to implementing support mechanisms and to incorporating additional appropriate levels of fault detection and fault tolerance at the processor, network, and operating system level are presented. A primary concern at these levels is balancing cost and performance against coverage and overall dependability. As these chapters demonstrate, low overhead, practical solutions are attainable and not necessarily incompatible with performance considerations. The section on innovative compiler support, in particular, demonstrates how the benefits of application specificity may be obtained while reducing hardware cost and run-time overhead.

The measurement of dependability attributes on real systems is a very time-consuming and costly affair, making analytical or simulation modeling the only viable solutions. Dependability of Networked Computer-based Systems explores reliability, availability and safety modeling of networked computer-based systems used in life-critical applications such as avionics, nuclear power plants, automobiles and chemical process industries. Dependability of Networked Computer-based Systems gives an overview of basic dependability modeling concepts and addresses new challenges in dependability modeling of networked computer-based systems, as well as new trends, their capabilities and limitations. It covers a variety of dependability modeling methods: stochastic processes, Markov and semi-Markov models, response-time distribution, stochastic Petri-net-based modeling formalisms, and Monte Carlo simulation models. Dependability of Networked Computer-based Systems provides students and researchers with a detailed overview of dependability models and analysis techniques. Practicing engineers will also find this text a useful guide to decision-making based on system dependability at the design, operation and maintenance stages.