Abstract

Any major research institution has a substantial number of computer systems on its campus, often in the scale of tens of thousands. Given that a large amount of scientific computing is appropriate for execution in an opportunistic environment, a campus grid is an inexpensive way to build a powerful computational resource. What is missing, though, is a model for making an informed decision on the cost-effectives of a campus grid. In this thesis, the author describes a model for measuring the costs and benefits of building a campus computing resource based on the institution’s existing investment in computing hardware.

For this study, the author calculates the usable capacity of a campus environ- ment, and based on TCO data for Purdue University’s campus grid, presents a model for calculating the base cost for a core-hour of computation in the campus grid.

Keywords

grid computing, condor, IT costs, cloud computing

Date of this Version

4-20-2011

Department

Technology

Department Head

Gary Bertoline

Month of Graduation

May

Year of Graduation

2011

Degree

Master of Science

Head of Graduate Program

Gary Bertoline

Advisor 1 or Chair of Committee

Jeffrey J. Evans

Advisor 2

Carol X. Song

Advisor 3

Kevin C. Dittman