Optimal Auction Design for WiFi Procurement

Abstract

The unprecedented growth of cellular traffic driven by web surfing, video streaming, and cloud-based services poses bandwidth challenges for cellular service providers. To manage the increasing data traffic, cellular service providers are experimenting the use of third-party WiFi hotspots to augment its cellular capacity. We develop an analytical framework to study the optimal procurement auction for WiFi capacity. Such an auction design is complicated by the fact that WiFi networks have much more limited spatial coverages compared with the cellular network. Neither a global auction that includes all WiFi hotspots nor multiple local auctions that include only hotspots in each local WiFi region is optimal. We find that the optimal mechanism is an integration of one global auction which includes hotspots from an endogeneously determined set of WiFi regions and many separate local auctions which are only held in the rest of the WiFi regions. To implement the optimal mechanism, we also provide an efficient algorithm whose computation complexity is of the order of the number of WiFi regions. Our work contributes to the literature by designing the optimal mechanism for a unique type of IT procurement auction problem which is a tight integration of economics and computational technology.

Keywords: WiFi offloading, procurement auction, mechanism design

JEL Classification: D44, L96, D86