The Dynamics of Price, Revenue and System Utilization

Srinivasan Jagannathan and  Kevin C. Almeroth

Department of Computer Science
University of California
Santa Barbara, CA 93106-5110
{jsrini,almeroth}@cs.ucsb.edu 
 
Content delivery is a growing enterprise in the Internet.  Critical to the
management of content delivery systems is understanding customer behavior,
its impact on the consumption of system resources and how together, these
affect revenue.  We believe that price dictates overall customer behavior
and that understanding the relationship between price and customer
behavior is the key to effective system management. To this end, we study
the impact of price on revenue and system utilization.  Our service model
is based on customers being able to refuse content based on their capacity
to pay and their {\em willingness} to pay the price quoted. We quantify
the effects of such customer behavior on revenue and system utilization.
We argue that charging a constant price based on the customers' capacities
to pay maximizes the expected revenue and formulate how to achieve this
maximum.  However, since customer behavior and characteristics are highly
varying and not known {\it a priori}, we develop an adaptive pricing model
which tracks user behavior as well as the arrival process to maximize
revenue. We validate it using simulation. Our simulation results indicate
that the adaptive pricing scheme generates nearly the same revenue as the
theoretical expectation under very dynamic workloads.