Tutorial: Architecting Fault Tolerant Systems

18th IEEE Int. Symposium on Software Reliability Engineering
                              ISSRE 2007
              November 6, Trollhättan, Sweden
              http://www.di.univaq.it/TutorialAFTS/
                         Call for Participation

This tutorial describes how the concepts of fault tolerance and software
architectures have been integrated so far in a new emerging domain
of architecting fault tolerant systems.

Fault tolerance (FT), being one of the four means for guaranteeing
dependability, aims to ensure the delivery of the correct services in
the presence of active faults. It is implemented by error detection and
subsequent system recovery. Software Architecture (SA)  represents
the first complete system description in the development life-cycle and
can help improving  the overall  system dependability, providing
a system blueprint that can be validated and that can guide all
phases of the system development. 

While typical SA specifications model only the normal behaviour of the
system, several approaches have been recently introduced for modelling
and analyzing fault tolerant software architectures that consider also
abnormal behaviours and define the architectural interconnections
between these two behaviour types.

This full day tutorial is structured into six parts:

1. Introduction to basic concepts in SA
2. A working example on SA: Insulin pump system (to be done
     with the attendees)
3. Architecture Description Languages: an Overview (focus on AADL)
4. Brief introduction to FT
5. Architecting Fault Tolerant Systems: the Survey
6. Case studies, tools, and examples (AADL, Coordinated Atomic
     Actions, Idealized FTCM and iC2C)

The value of this tutorial is in gaining deep understanding of the state
of the art in this emerging area.  This will be built on developing
a two dimensional view on the existing solutions: one dimension is
based on the traditional software architecture parameters and
another one on the typical fault tolerance related parameters.

The attendees of this tutorial will learn how to architect fault tolerant
systems and will become familiar with our approach to comparative
analysis of the existing architectures. They will be in a much better
position to understand current trends in the area and follow the ongoing
development. They will be prepared to apply the existing solutions
and, when necessary, to develop the new ones which are specific for
their domains, requirements and types of faults. Future trends in
architecting fault tolerant systems and open questions will be
discussed with the attendees.

The last part of this tutorial will present some case studies and
tools that are available. In particular, it will introduce specific
technologies for fault tolerance and, specifically, exception handling,
based on the Coordinated Atomic Actions, and illustrate how their
concepts can be expressed in the architectural models.


Lecturers:
H. Muccini, P. Pelliccione
	Dipartimento di Informatica
	University of L'Aquila, Italy
	[muccini,pellicci]@di.univaq.it
A. Romanovsky
	Center for Software Reliability
	Newcastle University, UK
	[log in to unmask]