A persuasive coin is a sufficiently unbiased source of randomness visible to sufficiently many processors in a distributed system. An algorithm is described for achieving a persuasive coin in the presence of an extremely powerful adversary where the number of rounds of message exchange among the processors is constant, independent of the number n of processors in the system as well as the number of faults, provided the total number of faulty processors does not exceed a certain constant multiple of $n/\log n$. As a corollary an $\Omega (n/\log n)$-resilient probabilistic protocol for Byzantine agreement running in constant expected time is obtained. Combining this with a generalization of a technique of Bracha, a probabilistic Byzantine agreement protocol tolerant of almost ${n / 4}$ failures with $O(\log \log n)$ expected running time is obtained.
[1]
Andrei Z. Broder,et al.
Flipping coins in many pockets (Byzantine agreement on uniformly random values)
,
1984,
FOCS.
[2]
Nancy A. Lynch,et al.
Impossibility of distributed consensus with one faulty process
,
1983,
PODS '83.
[3]
Nathan Linial,et al.
Collective coin flipping, robust voting schemes and minima of Banzhaf values
,
1985,
26th Annual Symposium on Foundations of Computer Science (sfcs 1985).
[4]
William Feller,et al.
An Introduction to Probability Theory and Its Applications
,
1967
.
[5]
Gabriel Bracha,et al.
An O(log n) expected rounds randomized byzantine generals protocol
,
1987,
JACM.