http://www.dagstuhl.de/13022

06. – 11. Januar 2013, Dagstuhl Seminar 13022

Engineering Resilient Systems: Models, Methods and Tools

Organisatoren

Nicolas Guelfi (University of Luxembourg, LU)
Maritta Heisel (Universität Duisburg-Essen, DE)
Mohamed Kaaniche (LAAS – Toulouse, FR)
Alexander Romanovsky (University of Newcastle, GB)

Koordinatoren

Elena Troubitsyna (Abo Akademi University, FI)

Auskunft zu diesem Dagstuhl Seminar erteilt

Dagstuhl Service Team

Dokumente

Dagstuhl Report, Volume 3, Issue 1 Dagstuhl Report
Motivationstext
Teilnehmerliste
Gemeinsame Dokumente

Summary

The Dagstuhl Seminar 13022 -- Engineering Resilient Systems: Models, Methods and Tools has brought together prominent researchers from different fields to discuss the problems of engineering resilient systems. The seminar was run in a highly interactive manner. The discussions were centered around the following topics:

  • defining resilience
  • resilience in modelling languages for requirement analysis and system design
  • resilience in implementation languages and frameworks
  • verifying resilience using testing, model checking and static analysis
  • assessing resilience using probabilist models
  • resilience mechanisms at architectural and implementation level

The concept of resilience has been introduced to capture the move towards a greater adaptability and flexibility. However, the notion of resilience is still a subject of debates. The seminar has discussed various proposed definitions and converged to defining resilience as dependability in presence of changes.

Over the last decades a remarkable progress has been achieved in engineering of highly dependable systems, i.e., the systems that can be justifiably trusted to provide critical services to a society. However, novel computing paradigms pose new scientific and technological challenges to the dependability field. To deliver critical services in a dependable way, the systems should smoothly adapt to changes. At the seminar, we had a dedicated session discussing the nature of changes. Among the proposed categories were

  • evolving user requirements
  • changing operating environment
  • unforeseen failure modes
  • scalability challenge

Modelling is the primarily vehicle driving development of resilient systems. However, system modelling area is still highly fragmented. The most acute problems are caused by

  • the gap between the requirements and models and
  • heterogeneity of models used to represent different aspects of system behaviour

Indeed, over the last few years the problem of poor flow-down of system requirements to software requirements has started to receive a proper attention. The vast majority of development relate the severe design problems with the flawed requirements and misunderstandings about what the software should do. Requirements tend to focus on describing nominal behaviour while omitting or poorly describing off-nominal conditions, safety constrains and fault tolerance mechanisms.

During the seminar we have brainstormed the examples of requirements that would be specific to resilient systems and tried to linked them with the modelling techniques.

While developing resilient systems the designers use dedicated models to reason about different (often antagonistic) aspects of system behaviour. Hence, the design space is inherently heterogeneous. On the one hand, specialised models provide the designers with expressive and powerful techniques to analyse various aspects of system behaviour. On the other hand, it becomes hard to obtain a holistic view on the system characteristics and analyse trade-offs between several potentially conflicting goals, define the mechanisms for adapting to volatile operating conditions and devise appropriate mechanisms for proactive fault tolerance.

We have discussed the advances in formal modelling of resilient systems and in particular proactive fault tolerance and adaptive fault tolerance mechanisms at various frameworks. We have reviewed the advances achieved in the area of formal modelling of resilient systems and brain-stormed the techniques leveraging an integration of various models to facilitate emergence of integrated modelling approaches.

Essentially, any design flow can be seen as a set of well-defined abstraction levels. The design flow should allow the designer to optimize design decision at each level and move freely between abstraction layers. At our seminar we discussed the principles of mapping abstract models onto architectural models and design implementation. We addressed the problem of achieving architectural plasticity and brain-stormed architectural patterns supporting adaptation as well as mechanisms guaranteeing adequate predictable system reaction on changes. A significant attention has also been paid to the methods and tools for resilience assessment.

Open Problems

Engineering resilient systems is a young research area. The participants of the seminar have agreed that often it is hard to distinguish a traditional dependability research from the resilience research. We have converged to the view that the system ability to scale, cope with changes and evolve emphasizes the resilience aspect.

It was also noted that the area of resilience engineering lacks a comprehensive reference guide that would allow the designers of resilient systems understand how various proposed methods and tools can facilitate design of resilient systems. The participants of the seminar has decided to work on such a book.

License
  Creative Commons BY 3.0 Unported license
  Maritta Heisel and Mohamed Kaaniche and Alexander Romanovsky and Elena Troubitsyna

Classification

  • Semantics / Formal Methods
  • Software Engineering
  • Verification / Logic

Keywords

  • Resilience
  • Dependability
  • Software engineering
  • Model-driven engineering
  • Formal methods
  • Verification
  • Fault tolerance
  • Requirements engineering

Buchausstellung

Bücher der Teilnehmer 

Buchausstellung im Erdgeschoss der Bibliothek

(nur in der Veranstaltungswoche).

Dokumentation

In der Reihe Dagstuhl Reports werden alle Dagstuhl-Seminare und Dagstuhl-Perspektiven-Workshops dokumentiert. Die Organisatoren stellen zusammen mit dem Collector des Seminars einen Bericht zusammen, der die Beiträge der Autoren zusammenfasst und um eine Zusammenfassung ergänzt.

 

Download Übersichtsflyer (PDF).

Publikationen

Es besteht weiterhin die Möglichkeit, eine umfassende Kollektion begutachteter Arbeiten in der Reihe Dagstuhl Follow-Ups zu publizieren.

Dagstuhl's Impact

Bitte informieren Sie uns, wenn eine Veröffentlichung ausgehend von
Ihrem Seminar entsteht. Derartige Veröffentlichungen werden von uns in der Rubrik Dagstuhl's Impact separat aufgelistet  und im Erdgeschoss der Bibliothek präsentiert.