A Measurement Study of Campus WiFi Networks Using WiFiTracer

Abstract

Highly dense and large-scale WiFi networks have been widely deployed in public areas to provide cost-effective high-speed wireless Internet access for mobile end users. This emerging practice has been leading to a severe spectrum usage overlap and channel interference between colocated WiFi networks. To understand the characteristics of highly dense WiFi networks, we conduct a measurement study of campus WiFi networks in this chapter. First, we instrument an Android App to sense WiFi access points (APs) to characterize WiFi networks in campus areas, including WiFi spectrum and channel usage, AP density, network distribution, and so on. Our measurement results demonstrate that a large number of WiFi APs have been widely deployed on campus, and about 80% of the total APs occupy the 2.4 GHz band, whereas the remainder part are the higher frequency 5 GHz APs, commonly used by public WiFi networks. The spectrum overlap and channel interference in the 2.4 GHz band is much more severe than that in the 5 GHz band. Then, extra WiFi connection measurements are conducted at selected areas with well-deployed campus WiFi networks, to understand WiFi connection characteristics while pedestrians are moving around in the coverage of the WiFi networks. By harvesting data from voluntary Android smart phone users, the connection setup time composed of Authentication–Association (AA) time, handshake time, and IP acquisition time is found to be generally affected by various factors, such as AP density, RSSI levels, etc. To achieve load balancing with reduced interference and higher WiFi network performance, this field measurement study may provide guidelines to design the next generation software-defined WiFi networks.