Course code TA02
Course title Software reliability
Institution Ecole Nationale Supérieure de Techniques Avancées
Course address ENSTA ParisTech – Palaiseau
City Paris
Minimum year of study 3rd year
Minimum level of English Good
Minimum level of French None
Key words

software reliability, operational environment fault removal, measurement, trend analysis

Language English
Professor responsible Florin POPENTIU
Telephone 01 45 81 78 19
Fax 01 45 81 31 19
Participating professors Florin POPENTIU, "UNESCO Chair" Department in Information Technologies, University of Oradea (Romania)/The Danish Technical University, DTU Informatics
Number of places Minimum: 10, Maximum: 30, Reserved for local students: 0

Motto: “Prediction is very difficult especially of the future” ( Niels Bohr)
The objective of this course is to answer the following questions:
What is software reliability?
Why are the statistical methods necessary?
How do you measure and predict the software reliability?
A computer is a deterministic machine - why can’t we predict when it will fail next ?
If software is such a problem why not build it in hardware?

There is evidence that defects have their origin in design errors. It becomes difficult or impossible to ensure that software contains no faults. The software reliability is currently a very sensitive area in telecommunications for example the introduction of new services.
The course presents opportunities in the field of prediction of software reliability and the tools allowing to characterize the accuracy and quality forecasts.
Various methods and techniques that we approach based on collected data: the software reliability growth models, statistical tests, among which trend tests (graphic and statistics methods). 

The course is focused on practical applications using software reliability toolkits on real world projects. 

Programme to be followed

Day 1: Key features of software systems; Trustworthy software; Software Forensics.
Day 2: Operational profiling and reliability modelling;  Measuring software reliability.
Day 3: Models for analysis of the software reliability growth; Reliability of Web services.

Day 4: Reliability assessment; Evaluation of software reliability predictions.
Day 5: Accuracy and quality of forecasts; Capabilities and limitations; Unanswered Questions; Case studies. 

Exercises: The students will experience teamwork. Project: Development of a software project by teams of students (usually five of them).

Methodological References Alternation of the exposed paper based on the course support under electronic format with the involvement of the students into debates. The copy of transparencies on CD-ROM and the Web pages: . Also the support of the course is accompanied by video illustrations and case studies with software tools. 

Prerequisites Basic knowledge in  statistics.
Course exam

Exam based on a mini-project programmed during the computer based sessions