Thursday, March 17, 2005
12:00 pm - 1:00 pm
Wean Hall, Room 8220
University of Pittsburgh
Prescience in Retrospect
With storage systems, and arguably any system that requires the transfer
of data, bandwidth can be overcome but latencies can only be avoided.
To avoid access latencies, the prediction of future accesses is attempted
(and often achieved) in one form or another (and not all obvious). Analyzing
access behavior and producing models of future accesses can be applied
in many different forms. This talk will discuss such access prediction,
a selection of application domains as time permits, and some of the more
common fallacies regarding the application of prediction in data storage.
Dr. Amer received his Ph.D. from the University of California,
Santa Cruz, where he worked primarily on predictive data management. He
still collaborates with the Storage Systems Research Center at UCSC, and
his most recent work has focused on the applications of access prediction
to construct adaptive caching schemes, self-optimizing systems algorithms,
and the analysis and enhancement of storage device power management.
Seminar Info Contact:
or visit http://www.pdl.cmu.edu/SDI/