37th Annual IEEE/IFIP International Conference on Dependable Systems and Networks (DSN'07) 2007
DOI: 10.1109/dsn.2007.62
|View full text |Cite
|
Sign up to set email alerts
|

Measuring Availability in Optimistic Partition-Tolerant Systems with Data Constraints

Abstract: Replicated systems that run over partitionable environments, can exhibit increased availability if isolated partitions are allowed to optimistically continue their execution independently. This availability gain is traded against consistency, since several replicas of the same objects could be updated separately. Once partitioning terminates, divergences in the replicated state needs to be reconciled. One way to reconcile the state consists of letting the application manually solve inconsistencies. However, th… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1
1

Citation Types

0
4
0

Year Published

2008
2008
2018
2018

Publication Types

Select...
4
2
1

Relationship

1
6

Authors

Journals

citations
Cited by 8 publications
(4 citation statements)
references
References 29 publications
0
4
0
Order By: Relevance
“…Regarding replication protocols, QB S should be discarded, since it was defined for achieving strong consistency and adopts the primary partition model [15] in case of disconnection. Apparently, PC shares the same problem, but it has an important difference: the primary copy is an inherent leader process and behaves as a group representative for state transference once the network partition disappears [20], assuming that there is a primary process per partition [5]. Thus, PC becomes the best choice in this regard.…”
Section: Minimal Convergence Effort (Mce): There Is a Tie Between Allmentioning
confidence: 99%
“…Regarding replication protocols, QB S should be discarded, since it was defined for achieving strong consistency and adopts the primary partition model [15] in case of disconnection. Apparently, PC shares the same problem, but it has an important difference: the primary copy is an inherent leader process and behaves as a group representative for state transference once the network partition disappears [20], assuming that there is a primary process per partition [5]. Thus, PC becomes the best choice in this regard.…”
Section: Minimal Convergence Effort (Mce): There Is a Tie Between Allmentioning
confidence: 99%
“…The replication service in the CORBA implementation consists of a replication manager, a replication protocol called P4Log (see [4,7]) and a reconciliation protocol (CS) as shown in Figure 6. The figure also shows the control flow of a client invocation.…”
Section: Replication Supportmentioning
confidence: 99%
“…Simulations showed that the drawback of the 'non-availability' assumption can be severe in some scenarios; namely the time taken to reconcile could be long enough so that the non-availability of services during this interval would be almost as bad as having no degraded service at all (thereby no gain in overall availability). An early evaluation [4] of the continuous service protocol in a simulated environment (J-Sim) acted as a proof of concept. This paper extends the evaluation by embedding the reconciliation and replication mechanism in a CORBA platform and performance testing in a physical network.…”
Section: Introductionmentioning
confidence: 99%
“…This solution is similar in concept to some approaches used in reconciling processes for partitionable systems [7].…”
Section: Mobile Approachmentioning
confidence: 91%