A novel predicted replication strategy in cloud storage

Data replication is widely used in cloud storage and data grid to improve the parallel service efficiency and the performance of system, which can promote the file availability and system load balancing, reducing the response time with multiple copies. But high volume of big data gives a new, enormo...

Full description

Saved in:
Bibliographic Details
Published in:The Journal of supercomputing 2020-07, Vol.76 (7), p.4838-4856
Main Authors: He, Li, Qian, Zhicheng, Shang, Fengjun
Format: Article
Language:English
Subjects:
Cloud computing
Compilers
Computer Science
Data replication
Data storage
Interpreters
Processor Architectures
Programming Languages
Response time
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Data replication is widely used in cloud storage and data grid to improve the parallel service efficiency and the performance of system, which can promote the file availability and system load balancing, reducing the response time with multiple copies. But high volume of big data gives a new, enormous and rigorous challenge to data storage and business access of cloud storage, specially to the quality of cloud services. In this paper, a novel dynamic predicted replication strategy (DPRS) combining with the access frequency of files and prediction method is proposed to predict the future access of each file and calculating the optional number of replicas based on the real access and future access periodically. The experiment results show that DPRS can availably decrease the response time of a file request and reduce the additional cost of the cloud storage system simultaneously.
ISSN:0920-8542
1573-0484
DOI:10.1007/s11227-018-2647-4