|
Scott Clearwater and Bernardo A. Huberman
Information Dynamics Laboratory, HP Labs
Abstract
Since usage patterns of information technology within organizations can be
bursty, the peak demand for IT resources can at times exceed the
installed capacity within the enterprise. If providers of such peak
capacity emerge, as was the case for electricity and natural gas, the
problem arises as to how to efficiently provide and price such peak
demand.
We present a swing option
mechanism that allows for the efficient pricing of IT resources
ranging from CPU usage to storage and bandwidth. This mechanism allows
users to buy the right but not the obligation to
future peak use. A statistical simulation tool allows the users to price these swings
according to their own utilization patterns and to recover some
of their costs if the options are not exercised. The provider in turn
exploits its ability to statistically multiplex its resources to
price peak usage. The use of these swing options serves as an
incentive to the users to accurately forecasts of their own needs,
thus leading to more efficient utilization of the
provider’s resources.
To be presented at the 11th
International Conference on Computing in Economics, June 23 2005,
Full paper in PDF format
| 
Printable version
