What's missing here is an overview of randomized protocols. For example, the "Lamport papers" deal to a large extent with how to achieve consensus in the presence of faults.
He never mentions anywhere in those papers the existence of randomized solutions to these. In practice they are often easier to implement and are only randomized in so far as reaching agreement with probability 1. Which I'm happy with :-).
A good survey is www.cs.yale.edu/~aspnes/randomized-consensus-survey.pdf
He never mentions anywhere in those papers the existence of randomized solutions to these. In practice they are often easier to implement and are only randomized in so far as reaching agreement with probability 1. Which I'm happy with :-).
A good survey is www.cs.yale.edu/~aspnes/randomized-consensus-survey.pdf