Basics on performance, reliability and validation. Fundamentals of probability theory and combinatorial methods. Discrete time and continuous time Markov processes. Stochastic Petri Nets. Modeling formalisms: PN and extensions. Supporting tools for modeling. Fundamentals of measurement theory and its application. System testing: functional, fault injection, robustness. Planning of experimental campaigns. Dependability benchmarking. Supporting tools for experimental evaluation and monitoring
Reference book (in Italian): “L’Analisi Quantitativa dei Sistemi Critici”, a cura di Andrea Bondavalli, Società Editrice Esculapio, ISBN 978-88-7488-431-5, Prima edizione, 286 pagine, Marzo, 2011.
Additional teaching material (in English):
- Slides presented during the course.
- Papers published in Journals and in Conference proceedings.
Learning Objectives
The course provides students with the notions required to understand aspects of quantitative analysis and of quality of the service offered by computing systems, with emphasis on the analysis of performances and reliability.
Knowledge and understanding - At the end of the course students will learn how to correctly approach the problem of the quantitative analysis of computing systems, exploiting the basic principles necessary to understand the aspects of quantitative assessment and of quality of service, and deepening some analysis techniques both based on stochastic models and based on experimental campaigns.
Applying knowledge and understanding - At the end of the course students will learn how to select the most appropriate quantitative analysis methodology for the system under examination. They will also be able to define stochastic models for the targeted systems, and solve them also with the support of automatic tools. Finally, they will be able to design and perform experimental measurements on real systems or prototypes.
Prerequisites
None
Teaching Methods
CFU: 9
Total hours of the course: 270
Hours reserved to private study and other individual formative activities: 198
Prof. Andrea Bondavalli, by appointment.
Contact professor by phone number or e-mail (andrea.bondavalli@unifi.it).
DiMaI, Universita' di Firenze, Viale Morgagni, 65 - 50134 Firenze
Tel. 055 2751481
Dott. Paolo Lollini, by appointment.
Contact professor by phone number or e-mail (lollini@unifi.it).
DiMaI, Universita' di Firenze, Viale Morgagni, 65 - 50134 Firenze
Tel. 055 2751486
Dott. Tommaso Zoppi, by appointment.
Contact professor by phone number or e-mail (tommaso.zoppi@unifi.it).
DiMaI, Universita' di Firenze, Viale Morgagni, 65 - 50134 Firenze
Tel. 055 2751488
Type of Assessment
The Assessment consists of two parts: Development of a project (either on modeling or on experimental evaluation), and an oral interview. Conditional to the oral interview is the acceptance of the project.
The (individual) project consists of:
1) applying stochastic modeling approaches or experimental evaluation techniques to selected case-studies, possibly agreed with the student.
2) writing a report on the performed activities, with the description of the obtained results. The report should address the following main points: i) description of the main encountered problems, ii) description of the applied analysis methodology (modeling or experimental evaluation), iii) description of the obtained results.
Students will have to submit the report and the developed models or sw. The project, once accepted, will remain valid for the entire academic year.
The oral interview will include the discussion of the project, and questions on the arguments covered during the course.
Course program
Basics on performance and reliability analysis of systems and on systems validation. Definitions of performances and reliability indicators. Rules for building and validating models. Fundamentals of probability theory. Combinatorial methods. Discrete time Markov processes. Continuous time Markov processes. Queue theory: the MM1 queue and its variants. Stochastic Petri Nets. Modeling formalisms: PN and extensions (e.g., SAN, DSPN). Automatic supporting tools. Fundamentals of measurement theory and its application. System testing: functional, fault injection, robustness. Planning of experimental campaigns. Dependability benchmarking. Supporting tools for testing and monitoring.