TY - GEN
T1 - Pattern-directed replication scheme for heterogeneous object-based storage
AU - Zhou, Jiang
AU - Xie, Wei
AU - Dai, Dong
AU - Chen, Yong
N1 - Publisher Copyright:
© 2017 IEEE.
PY - 2017/7/10
Y1 - 2017/7/10
N2 - Data replication is a key technique to achievedata availability, reliability, and optimized performance indistributed storage systems and data centers. In recent years, with the emergence of new storage devices, heterogeneousobject-based storage system, such as a storage system withthe co-existence of hard disk drives and solid state drives, have become increasingly attractive as they combine merits ofdifferent storage devices to deliver better promise. However, existing data replication schemes do not place data basedon heterogeneous device characteristics as well as consideringdistinct data access patterns. In this paper, we introduce anovel data replication scheme PRS to achieve efficient datareplication for heterogeneous storage systems. Different fromtraditional schemes, the PRS groups objects according todata access patterns and distributes replicas to heterogeneousdevices with their features. It uses a pseudo random algorithmto optimize replica layout by considering storage device performanceand capacity. The experimental results confirm thatPRS is a highly efficient replication scheme for heterogeneousstorage systems.
AB - Data replication is a key technique to achievedata availability, reliability, and optimized performance indistributed storage systems and data centers. In recent years, with the emergence of new storage devices, heterogeneousobject-based storage system, such as a storage system withthe co-existence of hard disk drives and solid state drives, have become increasingly attractive as they combine merits ofdifferent storage devices to deliver better promise. However, existing data replication schemes do not place data basedon heterogeneous device characteristics as well as consideringdistinct data access patterns. In this paper, we introduce anovel data replication scheme PRS to achieve efficient datareplication for heterogeneous storage systems. Different fromtraditional schemes, the PRS groups objects according todata access patterns and distributes replicas to heterogeneousdevices with their features. It uses a pseudo random algorithmto optimize replica layout by considering storage device performanceand capacity. The experimental results confirm thatPRS is a highly efficient replication scheme for heterogeneousstorage systems.
KW - Access pattern
KW - Data distribution
KW - Data replication
KW - Heterogeneous storage
KW - Objectbased storage
UR - http://www.scopus.com/inward/record.url?scp=85027450652&partnerID=8YFLogxK
U2 - 10.1109/CCGRID.2017.148
DO - 10.1109/CCGRID.2017.148
M3 - Conference contribution
AN - SCOPUS:85027450652
T3 - Proceedings - 2017 17th IEEE/ACM International Symposium on Cluster, Cloud and Grid Computing, CCGRID 2017
SP - 645
EP - 648
BT - Proceedings - 2017 17th IEEE/ACM International Symposium on Cluster, Cloud and Grid Computing, CCGRID 2017
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 17th IEEE/ACM International Symposium on Cluster, Cloud and Grid Computing, CCGRID 2017
Y2 - 14 May 2017 through 17 May 2017
ER -