A highly reliable and parallelizable data distribution scheme for data grids

Torkestani, Javad Akbari

doi:10.1016/j.future.2012.07.006

Cited by 5 publications

(1 citation statement)

References 41 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Since distributed computing comprises of physical and virtual images in a large network [1], neither the volume, size, location, time processing and storage resources are detailed and restricted [2,3]. The characteristics of distributed computing removes restrictions in using the system and traditional networks in providing its services to the user [1].…”

Section: Introductionmentioning

confidence: 99%

Novelty circular neighboring technique using reactive fault tolerance method

Noor

Zian

Rahim

et al. 2019

IJECE

View full text Add to dashboard Cite

The availability of the data in a distributed system can be increase by implementing fault tolerance mechanism in the system. Reactive method in fault tolerance mechanism deals with restarting the failed services, placing redundant copies of data in multiple nodes across network, in other words data replication and migrating the data for recovery. Even if the idea of data replication is solid, the challenge is to choose the right replication technique that able to provide better data availability as well as consistency that involves read and write operations on the redundant copies. Circular Neighboring Replication (CNR) technique exploits neighboring policy in replicating the data items in the system performs well with regards to lower copies needed to maintain the system availability at the highest. In a performance analysis with existing techniques, results show that CNR improves system availability by average 37% by offering only two replicas needed to maintain data availability and consistency. The study demonstrates the possibility of the proposed technique and the potential of deploying in larger and complex environment.

show abstract