A categorical approach to distributed systems expressibility and knowledge
Proceedings of the eighth annual ACM Symposium on Principles of distributed computing
The distributed firing squad problem
SIAM Journal on Computing
Knowledge and common knowledge in a distributed environment
Journal of the ACM (JACM)
Knowledge and common knowledge in a byzantine environment: crash failures
Information and Computation
Coordinated traversal: (t+1)-round Byzantine agreement in polynomial time
Journal of Algorithms
Fully Polynomial Byzantine Agreement for Processors in Rounds
SIAM Journal on Computing
Using knowledge to optimally achieve coordination in distributed systems
Theoretical Computer Science
Reaching Agreement in the Presence of Faults
Journal of the ACM (JACM)
Reasoning About Knowledge
Common knowledge and consistent simultaneous coordination
Distributed Computing
'Eventual' is earlier than 'immediate'
SFCS '82 Proceedings of the 23rd Annual Symposium on Foundations of Computer Science
Revisiting simultaneous consensus with crash failures
Journal of Parallel and Distributed Computing
Continuous consensus with ambiguous failures
ICDCN'08 Proceedings of the 9th international conference on Distributed computing and networking
| Hi-index | 0.00 |
The general omissions failure model, in which a faulty process may omit both to send and to receive messages is inherently more complex than the more popular sending omissions model. This fact is exemplified in tasks involving simultaneous decisions, such as the simultaneous consensus (SC) problem. While efficient polynomial protocols for SC that are optimal in all runs are known for the sending omissions model, they do not exists for general omissions. It has been shown that such a protocol must perform at least NP-hard computations (in the number of processes n) between rounds. In fact, the best previously known SC protocol that is optimal in all runs in this model performs PSPACE (in n) computations between rounds. The current paper closes this twenty-year old gap by presenting such an optimal SC protocol that performs PNP computations (polynomial-time computations using an oracle for NP; in fact, a constant number of accesses to the oracle are needed per round.) The result is based on a new characterization of common knowledge in the general omissions failure model.