Alternators in read/write atomicity

Authors:
Sandeep S. Kulkarni;Chase Bolen;John Oleszkiewicz;Andrew Robinson
Affiliations:
Department of Computer Science and Engineering, Michigan State University, East Lansing, MI 48824, USA;Department of Computer Science and Engineering, Michigan State University, East Lansing, MI 48824, USA;Department of Computer Science and Engineering, Michigan State University, East Lansing, MI 48824, USA;Department of Computer Science and Engineering, Michigan State University, East Lansing, MI 48824, USA
Venue:
Information Processing Letters
Year:
2005

Citing 8
Cited 2

A transformation of self-stabilizing serial model programs for asynchronous parallel computing environments

Information Processing Letters
Self-stabilizing systems in spite of distributed control

Communications of the ACM
Stabilization-preserving atomicity refinement

Journal of Parallel and Distributed Computing - Self-stabilizing distributed systems
A Timestamp Based Transformation of Self-Stabilizing Programs for Distributed Computing Environments

WDAG '96 Proceedings of the 10th International Workshop on Distributed Algorithms
Self-Stabilizing Local Mutual Exclusion and Daemon Refinement

DISC '00 Proceedings of the 14th International Conference on Distributed Computing
Mutual Exclusion Between Neighboring Nodes in an Arbitrary System Graph Tree That Stabilizes Using Read/Write Atomicity

Euro-Par '99 Proceedings of the 5th International Euro-Par Conference on Parallel Processing
The alternator

ICDCS '99 Workshop on Self-stabilizing Systems
Self-stabilizing atomicity refinement allowing neighborhood concurrency

SSS'03 Proceedings of the 6th international conference on Self-stabilizing systems

A uniform process alternator for arbitrary topologies

Journal of High Speed Networks
Self-stabilizing algorithms

Algorithms and theory of computation handbook

Quantified Score

Hi-index	0.89

Visualization

Abstract

The alternator problem requires that in legitimate states no two neighboring processes are enabled and between two executions of a process, its neighbors execute at least once. In this paper, we present a solution for the alternator problem that has the following properties: (1) If the underlying topology is arbitrary and the program is executed in read/write atomicity then it is stabilizing fault-tolerant, i.e., starting from an arbitrary state, it recovers to states from where its specification is satisfied, (2) If the underlying topology is bipartite and the program is executed in the concurrent execution model then it provides stabilizing fault-tolerance and maximal concurrency, (3) If the underlying topology is linear or tree then the program provides both these properties, and (4) The program uses bounded state if the network size is known. To our knowledge, this is the first alternator program that achieves these properties.