Fairness
Parallel program design: a foundation
Parallel program design: a foundation
Linearizability: a correctness condition for concurrent objects
ACM Transactions on Programming Languages and Systems (TOPLAS)
ACM Transactions on Programming Languages and Systems (TOPLAS)
Verification of a Distributed Summation Algorithm
CONCUR '95 Proceedings of the 6th International Conference on Concurrency Theory
ProCoS Proceedings of the Third International Symposium Organized Jointly with the Working Group Provably Correct Systems on Formal Techniques in Real-Time and Fault-Tolerant Systems
The design of a linearization of a concurrent data object
PROCOMET '98 Proceedings of the IFIP TC2/WG2.2,2.3 International Conference on Programming Concepts and Methods
Wait-free linearization with an assertional proof
Distributed Computing
Wait-free concurrent memory management by create and read until deletion (CaRuD)
Distributed Computing
| Hi-index | 0.00 |
Progress is investigated for a shared-memory distributed system with a weak form of fault tolerance that allows processes to stop and restart functioning without notification. The concept of bounded fairness is introduced to formalize bounded delay under the assumption that each family of related processes continuously contains at least one active member. This is a generalization of wait-freedom, and also of a finitary form of weak fairness. Several useful proof rules are stated and proved. In a system with bounded fairness, a wait-free process can be constructed by forming a new process in which processes from the various families are scheduled in a round robin way. The theory is applied to prove progress within bounded delay for a linearizing concurrent data-object in shared memory. The safety properties of this algorithm have been treated elsewhere.