Analysis of Multi-server Systems via
Dimensionality Reduction of Markov Chains

Takayuki Osogami

June, 2005

School of Computer Science
Carnegie Mellon University
Pittsburgh, PA 15213

Submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy

Thesis Committee:
Mor Harchol-Balter, Chair
Hui Zhang
Bruce M. Maggs
Alan Scheller-Wolf (Tepper School of Business, Carnegie Mellon University)
Mark S. Squillante (T. J. Watson Research Center, IBM Research)

Copyright © 2005 Takayuki Osogami

This research was supported by grant sponsorship from IBM Japan and IBM Tokyo Research Laboratory.

Keywords: Performance analysis, Multi-server system, Markov chain, Resource allocation, Cycle stealing, Priority, Task assignment, Phase type, Dimensionality reduction, Moment matching.