Work

A Methodology and Framework for Model-Driven Dependability Analysis of Critical Embedded Systems and Directions Towards Systems of Systems

Year: 2013
Type: dissertation
Abstract: In different domains, engineers have long used models to assess the feasibility \nof system designs; over other evaluation techniques modeling has the key \nadvantage of not exercising a real instance of the system, which may be costly, \ndangerous, or simply unfeasible (e.g., if the system is still under design). \nIn the development of critical systems, modeling is most often employed as \na fault forecasting technique, since it can be used to estimate the degree \nto which a given design provides the required dependability attributes, i.e., \nto perform quantitative dependability analysis. More in general, models are \nemployed in the evaluation of the Quality of Service (QoS) provided by the \nsystem, under the form of dependability, performance, or performability \nmetrics. From an industrial perspective, modeling is also a valuable tool \nin the Verification & Validation (V&V) process, either as a support to the process \nitself (e.g., FTA), or as a means to verify specific quantitative or qualitative \nrequirements. \n \nModern computing systems have become very different from what they used \nto be in the past: their scale is growing, they are becoming massively distributed, \ninterconnected, and evolving. Moreover, a shift towards the use of \noff-the-shelf components is becoming evident in several domains. Such increase \nin complexity makes model-based assessment a difficult and time-consuming \ntask. In the last years, the development of system has increasingly adopted the \nComponent-Based Development (CBD) and Model-Driven Engineering (MDE) \nphilosophies as a way to reduce the complexity in system design and evaluation. \nCBD refers to the established practice of building a system out of reusable \n“black-box” components, while MDE refers to the systematic use of models as \nprimary artefacts throughout the engineering lifecycle. Engineering languages \nlike UML, BPEL, AADL, etc., allow not only a reasonable unambiguous \nspecification of designs, but also serve as the input for subsequent development \nsteps like code generation, formal verification, and testing. One of the core \ntechnologies supporting model-driven engineering is model transformation. \nTransformations can be used to refine models, apply design patterns, and \nproject design models to various mathematical analysis domains in a precise \nand automated way. In recent years, model-driven engineering approaches have \nbeen also extensively used for the analysis of the extra-functional properties of \nthe systems. To this purpose, language extensions were introduced and utilized \nto capture the required extra-functional concerns. \nDespite several approaches propose model transformations for dependability \nanalysis, still there is not a standard approach for performing dependability \nanalysis in a MDE environment. Indeed, when targeting critical embedded \nsystems, the lack of support for dependability attributes, and extra-functional \nattributes in general, is one of the most recognized weaknesses of UML-based \nlanguages. Also, most of the approaches have been defined as extensions to \na "general" system development process, often leaving the actual process \nunspecified. Similarly, supporting tools are typically detached from the design \nenvironment, and assume to receive as input a model satisfying certain constraints. \nWhile in principle such approach allows not to be bound to specific \ndevelopment methodologies, in practice it introduces a gap between the design \nof the functional system model, its enrichment with dependability information, \nand the subsequent analysis. Finally, the specification of properties our of \ncomponents' context, which typically holds for functional properties, is much less \nunderstood for non-functional properties. \n \nThe work in this thesis elaborates on the combined application of the CBD \nand MDE philosophies and technologies, with the aim to automate dependability \nanalysis of modern computing systems. A considerable part of the work \ndescribed in this thesis has been carried out in the context of the ARTEMIS-JU \n“CHESS” project, which aimed at defining, developing and assessing a \nmethodology for the component-based design and development of embedded \nsystems, using model-driven engineering techniques. \nThe work in this thesis defines and realizes an extension to the CHESS \nframework for the automated evaluation of quantitative dependability properties. \nThe extension constitutes of: i) a set of UML language extensions, collectively \nreferred to as DEP-UML, for modeling dependability properties relevant for \nquantitative analysis; ii) a set of model-transformation rules for the automated \ngeneration of Stochastic Petri Nets (SPNs) models from system designs enriched \nwith DEP-UML; and iii) a model-transformation tool, realized as a plugin \nfor the Eclipse platform, concretely implementing the approach. After introducing \nthe approach, we detail its application with two case studies. \nWhile for embedded systems it is often possible, or even mandatory, to follow \nand control the whole design and development process, the same does not \nhold for other classes of systems and infrastructures. In particular, large-scale \ncomplex systems don’t fit well in the paradigm proposed by the CHESS project, \nand alternative approaches are therefore needed. Following this observation, \nwe then elaborate on a workflow for applying MDE approaches to support the \nmodeling of large-scale complex systems. The workflow is based on a particular \nmodeling technique, and a supporting domain-specific language, TMDL, which \nis defined in this thesis. After introducing a motivating example, the thesis \ndetails the workflow, introduces the TMDL language, describes a prototype \nrealization of the approach, and describes the application of the approach to \ntwo examples. We then conclude with a discussion and a future view on how \nthe contribution of this thesis can be extended to a comprehensive approach \nfor dependability and performability evaluation in a "System of Systems" context. \n \nMore in detail, this dissertation is organized as follows. Chapter 1 introduces \nthe context of the work, describing the main concepts related to dependability, \nand dependability evaluation, with a focus on model-based assessment. The \nfoundation of CBD and MDE approaches, the role of the UML language, and \nmain related work are instead discussed in Chapter 2. \nChapter 3 describes the CHESS project, and introduces the language extensions \nthat have been defined to support dependability analysis. Moreover, the \nchapter details the entire process that drove us to such extensions, including the \nelicitation of language requirements and the evaluation of existing languages \nin the literature. The model-transformation algorithms for the generation of \nStochastic Petri Nets are described in Chapter 4, while the adopted architecture \nfor the concrete realization of the analysis plugin is described in Chapter 5. \nChapter 6 describes the application of our approach to two case studies: of a \nmultimedia processing workstation and a fire detection system. \nThe need for a complementary approach for the evaluation of large-scale \ncomplex system is discussed in Chapter 7, with the aid of a motivating example \nof a distributed multimedia application. Chapter 8 describes our approach \nfor the automated assembly of large dependability models through model-transformation. \nThe thesis then concludes with an outlook on the relevance of the work \npresented in this thesis towards a System of Systems approach to the \nevaluation of large-scale complex systems.
Language: English
Open Access status: green
OpenAlex ID: W2720001662
Cites: 147
Cited by:
Related to: 10