An openflow-based wireless user management system
Abstract
The degradation of the end-to-end Quality of Service (QoS) on wireless users is becoming a common issue for IEEE 802.11 infrastructure-based networks in crowded areas. In response, network providers usually increase coverage and capacity to serve the ever-increasing simultaneous connections and traffic loads. However, wireless users still experience unacceptable QoS at the application layer (i.e.: video and audio streaming). In many cases, this issue is due to the nature of the IEEE 802.11 standard that makes wireless users select the target AP to camp on based on signal strength, regardless of traffic load supported by the AP at the moment. This thesis tackles the issue from a different perspective than previous works, by employing an SDN framework on an integrated wireless/wired environment. Thereby, we present the development and implementation of a system that performs user management by analyzing the network load from the Openflow statistics, as well as, the wireless information collected from the associated users. In order to analyse the behaviour of the proposed user migration algorithm, we evaluate the system under scenarios with different traffic load and user session duration. From the experiments, we observed that in several cases, wireless users get a considerable QoS improvement at the application layer when the system is activated. Based on the results, we present an analysis on how and when user migration in multi-AP IEEE 802.11 networks is most effective.