A Multi-dimension Measurement Study of a Large Scale Campus WiFi Network
Congcong Miao,Jilong Wang,Hui Wang,Jun Zhang,Weiwei Zhou,Shengchao Liu
DOI: https://doi.org/10.1109/lcn.2018.8638246
2018-01-01
Abstract:The growing trend of wireless devices and WiFi networks poses significant management challenges to network administrators. Characterizing WiFi user behavior and understanding WiFi network usage pattern are helpful to identify the management challenges so that network administrators could manage WiFi networks more efficiently. In this work, we collect comprehensive datasets, i.e. DHCP dataset, AAA dataset, SNMP dataset of ACs in a large campus WiFi network. We provide a detailed measurement study from multiple dimensions, i.e. server plane, temporal plane, spatial plane and traffic plane. We observe that the WiFi network under study is far from optimal. First, the phenomenon of IP waste is severe due to the isolation between DHCP server and AAA server. Second, current deployment of network infrastructure resources is based on network administrators' experience and it results in that theWiFi performance varies a lot across different areas. Furthermore, we also study the user behavior with different types of devices and in different kinds of buildings. Our observations indicate that the WiFi network could be improved and managed more efficiently from multiple dimensions. We believe that this measurement study is helpful for network administrators and researchers to understand more about large scale WiFi networks.