We study the consensus problem in a shared memory model where all processes are programmed alike, there is no global synchronization, it is not possible to simultaneously reset all parts of the system to a known initial state, and processes may be faulty. We present a consensus protocol for n processes which can tolerate up to ⌈n/2⌉ - 1 failures and which uses a single (2⌈1.5n-1⌉)-valued shared register.
Bibliographical noteFunding Information:
Correspondence to: G. Taubenfeld, AT&T Bell Laboratories, 28-408, 600 Mountain Avenue, Murray Hill, NJ 07974, USA. * This work was supported in part by ONR contract NOO014-89-J-1980, by the National Science Foundation under grants CCR-8405478 and IRI-9015570, by the Hebrew Technical Institute scholarship, by the Technion V.P.R. Funds - Wellner Research Fund, and by the Foundation for Re-search in Electronics, Computers and Communications, administered by the Israel Academy of Sciences and Hu-manities.
- Fault tolerance
- crash failures
- initialization failures
- shared memory