Computing Reviews
Today's Issue Hot Topics Search Browse Recommended My Account Log In
Review Help
Search
The constraints of magnetic versus flash disk capabilities in big data analysis
Regola N., Cieslak D., Chawla N.  ASBD 2012 (Proceedings of the 2nd Workshop on Architectures and Systems for Big Data, Portland, OR, Jun 9, 2012)4-9.2012.Type:Proceedings
Date Reviewed: Dec 28 2012

Banks, airlines, and major enterprises all rely on the strategic data analysis of huge customer and business datasets stored in data warehouses. Intelligent decision support systems are used to delineate the patterns of customer behavior and transactions for fine-tuning business conduct. However, systems equipped with sophisticated mathematical algorithms ought to provide fast responses to real-time queries. The literature on the subject includes evaluations of the performance of data and file migration on earlier computer systems that used disks and tapes [1,2]. But what storage devices should today’s systems use to store and provide quick access to queries from large legacy databases containing years’ worth of business intelligence? How should the read/write and transaction access times be improved on current and emerging storage devices?

The authors of this paper compare the efficiency of storing and retrieving query results from traditional redundant array of inexpensive disks (RAIDs) to that of new solid state flash disks. They designed experiments to investigate the input/output (I/O) performance of RAIDs and solid state disks. Not surprisingly, the authors found the performance of RAIDs impressive in sequential read/write operations, compared to solid state disks. The experimental results highlight the superiority of random read/write operations on the solid state disks over the throughput of RAIDs with similar I/O and transaction operations.

This paper provides clear results in the continuing debate over the factors that affect fast storage and the processing of transactions on traditional and modern storage devices. Those results demonstrate the need to develop new algorithms for minimizing I/O delays to speed up the processing of transactions. However, shouldn’t academic researchers and industry experts focus more research on developing fast block management algorithms [3,4] and advanced applications of flash disks in workload characterization [5]?

Information technology and computer engineering students might find this paper’s exploration of the limitations of RAIDs and flash disks useful for their senior capstone projects. Of course, it all depends on the student’s background, research interests, and industry experience.

Reviewer:  Amos Olagunju Review #: CR140785 (1304-0314)
1) Gavish, B.; Liu Sheng, O. R. Dynamic file migration in distributed computer systems. Communications of the ACM 33, 2(1990), 177–189.
2) Smith, A. J. Long term file migration: development and evaluation of algorithms. Communications of the ACM 24, 8(1981), 521–532.
3) Koltsidas, I.; Viglas, S. D. Flashing up the storage layer. In Proceedings of the VLDB Endowment VLDB Endowment Inc., 2008, 514–525.
4) Kgil, T.; Roberts, D.; Mudge, T. Improving NAND flash based disk caches. In Proceedings of the 35th Annual International Symposium on Computer Architecture IEEE Computer Society, 2008, 327–338.
5) Kim, H.; Ahn, S. BPLRU: a buffer management scheme for improving random writes in flash storage. In Proceedings of the 6th USENIX Conference on File and Storage Technologies USENIX Association, 2008, 239–252.
Bookmark and Share
  Reviewer Selected
Featured Reviewer
 
 
Performance Attributes (C.4 ... )
 
Would you recommend this review?
yes
no
Other reviews under "Performance Attributes": Date
Attributes of the performance of central processing units: a relative performance prediction model
Ein-Dor P., Feldmesser J. Communications of the ACM 30(4): 308-317, 1987. Type: Article
Jul 1 1988
Performance estimation of computer communication networks: a structured approach
Verma P., Computer Science Press, Inc., New York, NY, 1989. Type: Book (9789780716781837)
Jun 1 1990
Computer hardware performance: production and cost function analyses
Kang Y. Communications of the ACM 32(5): 586-593, 1989. Type: Article
Feb 1 1990
more...

E-Mail This Printer-Friendly
Send Your Comments
Contact Us
Reproduction in whole or in part without permission is prohibited.   Copyright 1999-2024 ThinkLoud®
Terms of Use
| Privacy Policy