Efficient Capacity Provisioning for Firms with Multiple Locations: The Case of the Public Cloud

Abstract

We analyze a model in which a firm with multiple locations chooses capacity and prices to maximize efficiency. We find that the firm provisions capacity in such a way that the expected fraction of demand that will be unfilled is lower in locations with greater expected demand. The firm also sets lower prices in larger locations. Finally, if a customer is indifferent between multiple locations, then it is more efficient to place this customer in a location with greater expected demand. These theoretical results are consistent with empirical evidence that we present from a major public cloud provider.