CDRS:an adaptive cost-driven replication strategy in cloud storage

doi:10.7523/j.issn.2095-6134.2011.6.009

Abstract

Abstract:

Considering the new problems in cloud storage environment, such as wide location, dynamic characteristic, pursuit of commercial interests, and trade-off between consistency and availability, we propose a dynamic cost-driven replication strategy (CDRS). By introducing cost in market mechanism and considering load balancing and the trade-off between consistency and availability comprehensively, this strategy operates adaptively on replication and achieves the goal of minimizing replication cost and maximizing replication profit. The experimental results show that CDRS has a better performance in both replication profit and load balancing.

Key words: cost-driven, geographic feature, adaptive, cloud storage

CLC Number:

TP393

XU Jing, YANG Shou-Bao, WANG Shu-Ling, LIU Xiao-Qian. CDRS:an adaptive cost-driven replication strategy in cloud storage[J]. , 2011, 28(6): 759-767.

References

[1] 周建.教你实施云存储 .计算机世界,2010,8. .http://www.qikan.com.cn/Article/jsjj/jsjj201016/jsjj20101630.html.

[2] Sun公司. 云计算架构介绍白皮书 .2009:6.

[3] Birrell A D, Levin R, Needham R M, et al. Grapevine: An exercise in distributed computing
[J]. Communications of the ACM, 1982, 25(4):260-274.

[4] Bonvin N, Papaioannou T G, Aberer K. Dynamic cost-efficient replication in data clouds //ACDC'09. Barcelona, Spain, 2009.

[5] Pinheiro E, Weber W D, Barroso L A. Failure trends in a large disk drive population //Proc of 5th USENIX Conference on File and Storage Technologies (FAST'07). SanJose, CA, USA, 2007.

[6] Gilbert S, Lynch N. Brewer's conjecture and the feasibility of consistent, available, partition-tolerant web services
[J]. ACM SIGACT News, 2000, 33(2).

[7] Vogels W, Eventually consistent
[J]. Communications of the Acm, 2009, 52(1):40-44.

[8] Saito Y, Shapiro M. Optimistic Replication
[J]. ACM Computing Surveys, 2005, 37(1):42-81.

[9] Korupolu M, Plaxton G, Rajaraman R. Placement algorithms for hierarchical cooperative caching //Proceedings of the l0th Annual Symposium on Discrete Algorithms.1999.

[10] Li B, Golin M, Italiano G, et al. On the optimal placement of web proxies in the internet //Proceedings of IEEE INFOCOM'99, 1999.

[11] Yu H F, Amin Vahdat. Minimal replication cost for availability //PODC 2002, July 21-24. Monterey, Califormia, USA, 2002.

[12] Rahman R M, Barker K, Alhajj R. Replica placement strategies in data grid
[J]. Grid Computing, 2008, 6:103-123.

[13] Zhou X, Lu X L, Hou M S, et al. A dynamic distributed replica management mechanism based on accessing frequency detecting
[J]. Operating Systems Review, 2004, 38(3).

[14] DeCandia G, Hastorun D, Jampani M, et al. Amazon's highly available key-value store //Proceedings of the 21st ACM Symposium on Operating Systems Principles.

[15] Chen C T, Hsu C C, Wu J J, et al. GFS: A distributed file system with multi-source data access and replication for grid computing //4th International Conference on Grid and Pervasive Computing. Geneva,2009.

[16] Bonvin N, Papaioannou T G, Aberer K. The costs and limits of availability for replicated services //ACDC'09, June 19. Barcelona, Spain, 2009.