Fault‐Tolerant Consensus
Hagit Attiya, Jennifer Lundelius Welch · Distributed Computing · 2004
This chapter introduces two processor failure models, crash and Byzantine. It presents a fundamental coordination problem called consensus. Tight bounds are given on the number of rounds and on the ratio of faulty to nonfaulty processors for solving consensus in synchronous systems. Both shared memory and message passing is considered for the asynchronous case, while only message passing is considered for the synchronous case.