Course plan
Autumn 2003
Period: This is an intensive version of the course developed for
the CUGS (National Graduate School in Computer Science), and will run
over four days in November 2003.
Examiner: Simin
Nadjm-Tehrani,
Laboratory
of Real-time Systems, Dept. of Computer and Information Science
(IDA).
Organisation: Day 1 starts with an overview, followed by in
depth lectures. There will be a reflection/question-answering session
in the middle or end of the day. At the end of Day 2 the participants
are expected to formulate three learning goals for deeper study. Day 3
will start with a seminar where each participants describes the
progress achieved on the learning goals, and problems encountered.
There will 1 lecture by Ulf Nilsson
on self-stablising algorithms towards the end of the course. An invited
talk by another guest lecturer is also in progress.
The duration of each lecture varies between 2-3 hours depending on
ongoing discussions.
Planned lectures/resource sessions:
Monday 3/11:
- Introduction to the course, basic notions in
fault-tolerance
(FT) and a short review of the area in the intersection of FT and
formal
analysis of systems. Slides
in
pdf
- Introduction to basic notions in distributed systems,
including local/global state, cut and mechanisms for broadcast,
overview
on replication models. Slides
in pdf
- Consensus, and the related problems, the main impossibility
result. Slides
in pdf
Tuesday 4/11:
- Unreliable failure detectors. Slides
in pdf
Monday 24/11:
- Seminar: Review of studied material
- Group communication, view synchronous broadcast. Slides
in pdf
- Partitions and group management. Slides
in pdf
Tuesday 25/11:
- Guest lecture by Ulf Nilsson: Self-stablising algorithms.
Slides
in pdf
- Seminar: Synchronous networks, Proof for Byzantine agreement
protocol, Slides
in pdf
Last Modified 03-11-25
Simin Nadjm-Tehrani