PhD course by

Carlos Budde, Ph.D., Assistant Professor at the University of Trento (https://webapps.unitn.it/du/en/Persona/PER0235360/Curriculumhttps://scholar.google.it/citations?user=EmOwGskAAAAJ&hl=it&oi=ao), will present a Ph.D. course on “Simulations for Systems Safety and Security”, developing
on methods for Rare Events Simulation and focusing on their application to Dynamic Fault Trees and Attack Trees.
The course will be delivered on-line, combining presentations with hands-on tool-based experience, in 4 lectures of 3 hours each, in the afternoon of Luly 10,11,12, and 13 (12 hours, 3Credits)


TITLE: “Simulations for Systems Safety and Security”

CONTENT: the course develops on methods for Rare Events Simulation, focusing on their application to Dynamic Fault Trees and Attack Trees, combining presentations with hands-on tool-supported experience

INSTRUCTOR: Carlos Budde, Ph.D., Assistant Professor at the University of Trento (https://webapps.unitn.it/du/en/Persona/PER0235360/Curriculumhttps://scholar.google.it/citations?user=EmOwGskAAAAJ&hl=it&oi=ao)

SCHEDULE: the course will be delivered in 4 lectures, on-line, in the afternoon of July 10,11,12, and 13

Simulations for Systems Safety and Security (SSSS):
Formal methods takes on RES for DFT and AT analysis

This course gives a quick'n dirty introduction to system model analysis for safety and security in the presence of rare events.
We will overview the international fault-tree model standard, used for reliability and availability computations of complex fault-tolerant processes. We will also establish a parallel to attack trees, and note the connections between the two.
Failure and attack probabilities in these models are often inferred empirically. This means that the usual Markovian approximations fall short, and instead simulation (aka "statistical model checking") is often employed to compute the desired metrics. We will see the IOSA formalism that allows for this in a compositional manner, while supporting arbitrary probability distributions as model behaviour.
However, since the systems are fault-tolerant by design, faults become rare events and special methods to speed up convergence are needed. For this last challenge we will overview the basics of rare event simulation, focusing on importance splitting techniques.
The course includes hands-on experiences where we will generate concrete IOSA models from FTs and ATs, and analyse rare-event properties using the FIG statistical model checker.