Affiliations: Department of Computer Science, Faculty of Exact and Applied Sciences, University of Oran, Ahmed Ben Bella, Algeria
Correspondence:
[*]
Corresponding author: Said Limam, Department of Computer Science, Faculty of Exact and Applied Sciences, University of Oran 1, BP. 1524, EL M'Naouer, Oran, Algeria. E-mail:said.limam@gmail.com
Abstract: As cloud computing is becoming increasingly popular,
cloud storage services attract more attentions for their high security and
availability with a low cost. One way to provide availability and
scalability is by the use of replication, which places multiple copies of
the data in different servers in order to survive to server failures
(availability) and to distribute the request load (scalability). On the
other hand, the existence of multiple data replicas introduces a challenge
which is how to maintain replicas consistent. Initially we proposed in this
paper a model based on an optimistic approach using versions of vectors
between the replicas to manage the consistency of replicas. In a second step
we improved the approach proposed by the integration of a convergence
mechanism and conflict resolution between the replicas that suits the Clouds
Computing. This mechanism is based on the different forms of negotiation
between Local Consistency Service (LCS). We made the modeling and
formulation of the proposed model. We also evaluated the performance of this
model. The results obtained by the simulation show the effectiveness of our
approach in term of execution time and masking divergence and
incompatibility situations.
Keywords: Consistency, replica, conflict resolution, cloud computing, system availability, CloudSim
