Auction-based Resource Allocation for Wireless Local Area Networks in Metropolitan Areas Bo GU*, Kyoko YAMORI** *, Sugang XU*, Yoshiaki TANAKA* *** *Global Information and Telecommunication Institute, Waseda University, Tokyo, 169-0051 Japan **Department of Business Administration, Asahi University, Gifu, 501-0223 Japan ***Research Institute for Science and Engineering, Waseda University, Tokyo, 162-0044 Japan bo.gu@akane.waseda.jp, kyamori@alice.asahi-u.ac.jp, xusugang@m.ieice.org, ytanaka@waseda.jp Abstract— In this work, we focus on using pricing as an incentive mechanism to encourage self-interested access points to participate in a public wireless network. Specifically, we employ a centralized model to aggregate all of the access points deployed by different businesses into a federated market, enabling wireless clients to select the cost-effective access service with the consistent quality. The proposed pricing scheme is based on a second-price auction protocol. In addition to the cost- performance concern on the wireless access service, a travel- expense is introduced as another concern to the wireless clients for the service selection among different access points. Finally, we show that the proposed scheme outperforms a fixed-rate pricing scheme in terms of network utilization and resource allocation efficiency via simulation. Keywords— Pricing, Wireless Local Access Network, Auction, Travel-expense, Quality of Service I. INTRODUCTION In the recent years, we are witnessing a tremendous gain in the popularity of IEEE 802.11 Wireless Local Area Network (WLAN). In the U.S., several companies have declared their intentions to build and support nationwide WLAN networks, with tens of thousands of hot spots for providing broadband IP connectivity. On the other hand, a majority of city dwellers have a broadband connection and a personal access point at home. In dense metropolitan areas such as San Francisco, there is sufficient density of access points to achieve near- ubiquitous WLAN by sharing access amongst residents. Owners of WLANs would share their networks with the public if they could be adequately compensated. In this sense, pricing could be used as an incentive mechanism to encourage self-interested access points to share their networks [1]. In this work, we propose a novel pricing scheme under which the overall payment charged grows proportionally with the time the client gains access initially, and ceases when he no longer intends to connect. The proposed pricing scheme is based on a second-price auction mechanism, which is best known for its high efficiency in terms of resource allocation. In the proposed scheme, in addition to the cost-performance concern on the wireless access service, a travel-expense is introduced as another concern to the wireless clients for the service selection among different access points. The motivation of the work presented in this paper is to put forward a simple pricing scheme that (i) provides the right incentives for both access points and clients to follow the scheme and not to cheat so that the order in this federated market can be maintained spontaneously; (ii) improves network utilization and social welfare by introducing a second-price auction mechanism and enabling a mechanism for wireless clients to select nearby access points based on price concerns. II. SYSTEM MODEL We consider that multiple basic service sets (BSS) exist simultaneously in a certain area, such as a park or a community. Each BSS consists of an access point, and wireless clients who reach the access point directly. All access points in our model are registered with a trusted central authority - aggregator. The aggregator attaches its brand name to access points, so as to ensure that a consistent Quality of Service (QoS) is offered among the access points deployed by different businesses. Wireless clients are registered with the aggregator as well. By using a single account, a client can get an access service from any one of the federated access points, which have the same brand name of the aggregator. The aggregator can be distributed, as certification companies are, to avoid being a bottleneck. A. Quality of Service There are many factors that can influence the QoS, such as the number of clients coexisting, the power output, the type of antenna, the density of the building environment and the distance between a client and an access point [2]. To simplify the model in the study on pricing scheme, here we assume that the QoS provided by a branded access point is merely determined by the bandwidth allocated [3]. In this work we assume that, with an agreement between the aggregator and each access point, the same amount of dedicated bandwidth is offered among all the federated access points such that a consistent QoS is provided by different businesses. B. Web Browsing Utility When the allocated bandwidth stays unchanged, the client is considered to have a web browsing utility function [4]: his ISBN 978-89-5519-162-2 470 Feb. 19~22, 2012 ICACT2012