Affiliations: [a] University of Sciences and Technology, Mohammed Boudiaf, Oran, Algerie | [b] Computer Science Department, ISAMM, University of Manouba, Tunis, Tunisia | [c] University of Lyon, Lyon, France
Correspondence:
[*]
Corresponding author: Mohammed Rebbah, University of Sciences and Technology, Mohammed Boudiaf, El Mnaouar, BP 1505, Bir El Djir 31000 Oran, Algerie. E-mail: [email protected]
Abstract: Grid reliability analysis and modeling are not easy tasks because of the complexity and large scale of the system. For some grid services which have large subtasks requiring time-consuming computation, the reliability of grid service could be rather low. To resolve this problem, RNFR and LNFR are proposed in literature. This paper describes the property of the grid computing systems and analyzes the grid program and service reliability. In this paper, we propose a fault tolerance mechanism, namely LRNFR, based both on RNFR and LNFR in a complementary manner with a periodic migration. We select the most reliable node among the set of Neighboring Collaborators. The migration process is divided by the MDTTR of the failed node until its repair or the migration of all its programs. Experiments demonstrate the performance of the proposed approach.
Keywords: Large scale systems, grid computing, fault tolerance, process migration, service reliability
