Revisiting Join Site Selection in Distributed Database Systems
New characteristics for e-commerce applications, such as highly distributed data and unpredictable system nature, require us to revisit query processing for distributed database systems. As join operations involve relations over multiple sites, the site to carry out the join operation can have a significant impact on the performance. In this paper, we propose a new join site selection method. Our method is different from traditional methods in the following ways. First, the candidate sites are not restricted to the operand sites or query site as adopted by most database systems. Second, it considers dynamic system properties such as available bandwidth and system loads. We also conducted some experiments to offer a comparison between our method and the traditional one. The results show the advantage of our method under various system statuses.
KeywordsSite Selection Cost Model Database Server Server Load Candidate Site
Unable to display preview. Download preview PDF.
- 1.Stocker, K., Kossmann, D., Braumandl, R., Kemper, A.: Integrating Semi-Join-Reducers into State of the Art Query Processors. In: ICDE 2001, pp. 575–584 (2001)Google Scholar
- 2.Ye, H., Kerhervé, B., Bochmann, G.v., Oria, V.: Pushing Quality of Service Information and Requirements into Global Query Optimization. In: The Seventh International Database Engineering and Applications Symposium (IDEAS 2003), Hong Kong, China, July 16–18 (2003)Google Scholar
- 4.Cornell, D.W., Yu, P.S.: Site Assignment for Relations and Join Operations in the Distributed Transaction Processing Environment. In: ICDE 2001, 100–108 (1988)Google Scholar
- 5.Selinger, P.G., Adiba, M.: Access path selection in distributed data base management systems. In: Proc. of the International Conference on Data Bases, pp. 204–215 (1980)Google Scholar
- 6.Ye, H.: Integrating Quality of Service Requirements in a Distributed Query Processing Environment, Ph.D thesis, University of Montreal (May 2003)Google Scholar
- 7.National Laboratory for Applied Network Research, http://www.nlanr.net/