Price-based Congestion-Control in Wi-Fi Hot Spots

Wireless networks are now proliferating due to the success of the IEEE 802.11b protocol, also known as "Wi-Fi" (Wireless Fidelity). A Wi-Fi network is characterized by a set of base stations (also called access points) placed throughout the environment and connected to the traditional wired LANs. This technology allows nomadic users a broadband access to the Internet if they are in the transmission range of an access point. A new business model, named Wi-Fi Hot Spots, is now emerging to exploit the potentialities of this technology. A hot spot is a "critical" business area, e.g., airports, stations, hotels, where users can have wireless access by subscribing a contract with the hot spot operator, or with a wireless Internet service provider (WISP). Due to the random access nature of the Wi-Fi technology, if the number of users connected to the same access point increases, the QoS experienced may quickly degrade. This generates complains from the users that, as a consequence, may change their WISP. In order to be competitive, a Wi-Fi hot spot operator needs simple and effective mechanisms to control the congestion therefore guaranteeing the QoS, and (at the same time) maximizing his/her revenues. In this paper we present and evaluate a price-based policy for the access control in a Wi-Fi hot spot. Our policy, named Price-based Congestion Control (PCC), controls the hot spot traffic by dynamically determining the access cost as a function of the current load in the hot spot. We develop a theoretical framework to compute for any load condition the access cost to maintain the hot spot in its optimal operating point, for any load condition. The effectiveness and robustness of the PCC policy has been evaluated by simulating a Wi-Fi hot spot. Both in saturated and not- saturated conditions the PCC policy provides a better channel utilization than the legacy Wi-Fi policy.