Abstract
This lecture presented the consequences of questioning an assumption often underlying distributed algorithms : the number of machines involved in distributed computing is known in advance. This static case was contrasted with the dynamic case, in which the set of machines participating in the calculation is not known in advance.
Here are the main topics covered during this lecture :
- " Group membership " ;
- " From synchronism to virtual synchronism " ;
- " A reliable broadcast that ends " ;
- " The importance of uniformity
