Abstract
In shared-nothing environments, data is typically declustered and indexed across the system processing elements (PEs) to achieve efficient processing. However access patterns are inherently dynamic and skewed, thus, data reorganization based on the data access history (heat) is essential and should be done online. While the data is being reorganized, indexes need to be modified too, therefore, reorganization should additionally deal with the index modification. Based on minimization of index modification, we propose a data reorganization technique over a shared-nothing parallel system. By finding the exact work that should be done, the technique can smoothly balance a given heat across the PEs as fast as possible, if it is required. By tuning its parameters, it can cover a wide range of balancing requirements. We evaluate its performance through simulation studies. Its effectiveness is clarified quantitatively.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
Similar content being viewed by others
References
Achyutuni, K. J., Omiecinski, E., Navathe, S. B.: Two techniques for On-line Index Modification in Shared Nothing Parallel Databases. Procs ACM SIGMOD (1996)
Boral, H., et al: Prototyping Bubba, a Highly Parallel Databse System, IEEE Trans. On Knowledge and Data Eng., Vol. 2, No. 1, March (1990)
Copeland, G., Alexander, W., Boughter, E., Keller, T.: Data Placement in Bubba. Proc. of ACM SIGMOD Conference, pages 99–108, (1988)
DeWitt, D.J. and Gray, J.: The Future of High Performance Database Systems. Communication of ACM, 35(6), 85–98, (1992)
Lee, M. L., Kitsuregawa, M., Ooi, B.C., Tan, K, Mondal, A.: Towards Self-Tuning Data Placement in Parallel Database Systems, Proc. ACM SIGMOD pages 225–236 (2000).
Ozsu M., Valduriez., P.: Principles of Distributed Database Systems, Prentice Hall, (1991)
Salzberg, B., A. Dimock. Principles of transaction-based on-line reorganization. Procs. of the 18th Inter. Conf. on VLDB, pages 511–520, (1992)
Scheuermann, P., Weikum, G., Zabback, P., Adaptive Load Balancing in Disk Arrays. Proceedings of the 4th Inter. Conf. FODO, (1993)
Seeger B. and Larson P. Multi-Disk B-trees. ACM SIGMOD Conf.1991, 436–445
Tamura, T., Oguchi, M., Kitsuregawa, M.: Parallel Database Processing on a 100 Node PC Cluster: Case for Decision Support Query Processing and Data Mining. Proc. Of SC97: High Performance Networking and Computing, (1997)
Yokota, H., Kanemasa, Y., Miyaazaki, J.. Fat-Btree: An Update-Conscious Directory Structure. Procs. of IEEE the 15th IEEE Conf. on Data Engineering, pp. 448–457, (1999)
Zou C., Salzberg, B.: On-Line Reorganization of Sparsely-Populated B+ Trees. Procs. ACM, pages 115–124, (1996)
Valduriez, P.: Parallel Database Systems: Open Problems and New Issues, Distributed and Parallel Databases 1, No. 2, 137–165, Kluwer Academic Publishers, Boston, MA (1993).
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2000 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Feelifl, H., Kitsuregawa, M., Ooi, BC. (2000). A Fast Convergence Technique for Online Heatbalancing of Btree Indexed Database over Shared-nothing Parallel Systems. In: Ibrahim, M., Küng, J., Revell, N. (eds) Database and Expert Systems Applications. DEXA 2000. Lecture Notes in Computer Science, vol 1873. Springer, Berlin, Heidelberg. https://doi.org/10.1007/3-540-44469-6_79
Download citation
DOI: https://doi.org/10.1007/3-540-44469-6_79
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-67978-3
Online ISBN: 978-3-540-44469-5
eBook Packages: Springer Book Archive