The Choice Coordination Problem requiresn asynchronous processes to reach a common choice of one out ofk possible alternatives. Processes communicate viak shared variables. Up tot, t<n, of the processes may fail to operate by suddenly quitting the protocol. Rabin (1982) presented lower and upper bounds for the extreme caset=n−1. We present deterministic and randomized algorithms for arbitraryt using an alphabet of sizeO(t2). A semi-synchronous model is also studied. A reduction to a consensus problem proves the necessity to assume some powerful atomic shared-memory operations.
Key wordsChoice coordination Probabilistic versus deterministic algorithms Resilient algorithms
Unable to display preview. Download preview PDF.
- Ben-Or, M (1983) Another Advantage of Free Choice. Proc 2nd ACM Symp of Principles of Distributed Computing, 1983Google Scholar
- Fischer, MJ, Lynch, NA, Paterson, MS (1985) Impossibility of Distributed Consensus with One Faulty Process. J ACM 32:374–382Google Scholar
- Herlihy, MP (1988) Impossibility and Universality Results for Wait-Free Synchronization. Proc 7th ACM Symp of Principles of Distributed Computing, 1988, pp 276–290Google Scholar
- Rabin, MO (1982) The Choice Coordination Problem. Acta Inf 17:121–134Google Scholar