Power and performance optimization at the system level
Valentina Salapura, Randy Bickford, et al.
CF 2005
This paper considers the performance of a hypothetical multimicroprocessor backend database machine relative to that of a mainframe database system. The class of workloads considered is high volume transaction processing. Analytic queueing models of the two architectures are constructed and used in parametric performance studies. The performance sensitivities with respect to transaction complexity, the amount of overhead required to implement the distributed database function, the amount of skew in the data access pattern, and the buffer miss ratio are determined. These performance sensitivities are determined for technology available in both 1986 and for that projected for the early 1990's. © 1988, IEEE. All rights reserved.
Valentina Salapura, Randy Bickford, et al.
CF 2005
Narasimha R. Adiga, Matthias A. Blumrich, et al.
IBM J. Res. Dev
Paul Glasserman, Philip Heidelberger, et al.
Journal of Derivatives
David M. Nicol, Philip Heidelberger
SIGMETRICS 1993