Alma Riska


Department of Computer Science
Williamsburg, Virginia 23187-8795
E-mail: riska at cs dot wm dot edu
Curriculum Vitae: PDF
Research Statement:PDF
Teaching Statement:PDF
My work is centered around advanced data mining, analysis, and modeling, to devise techniques that improve data reliability, availability, and consistency, resource management, and performance of high performing data-centric systems. I focus on identifying metrics that accurately and compactly capture specific aspects of a system operation and incorporate them into resource management policies such that the systems adapt seamlessly to changes in system operation.
I completed my PhD at the Computer Science Department of the College of William & Mary in December 2002. The thesis of my PhD was that careful workload characterization and accurate modeling assists systems analysis both off- and on-line for higher adaptivity in today's dynamic operational environment. In my PhD, I extended and developed a new aggregation-based methodology, called ETAQA, that efficiently solves Markov processes of M/G/1 and QBD type. These processes were used to model and analyze clustered Web servers by capturing accurately salient characteristics of the Internet traffic, such as variability and burstiness. This analysis resulted in new adaptive load balancing policies in clustered Web servers farms.