Course goals

Autumn 2003

- To get an insight into the history of development and importance of fault-tolerance algorithms for distributed systems.

- To understand what are the major classifications, concepts and terms which define classes of distributed systems, fault-tolerance problems therein and solutions to some of these problems: Elements of robust algorithms, in particular concensus and broadcast algorithms, group mechanisms and stablising algorithms.

- To understand the underlying failure models for which strong results are available.

- To understand major limitations of achieving fault-tolerance with the help of robust algorithms in an asynchronous setting.

- To study a well-known problem that is solvable within a
synchronous
setting: Byzantine agreement.

- To get an overview of stablising algorithms and to study examples of such algorithms.

Last Modified 03-11-03

Simin Nadjm-Tehrani