Date Published: January 16, 2019
Publisher: Public Library of Science
Author(s): Mun-Suk Kim, Yena Kim, SeungSeob Lee, SuKyoung Lee, Nada Golmie, Chi-Tsun Cheng.
The current commercial access point (AP) selection schemes are mostly based on received signal strength, but perform poorly in many situations. To address this problem, a number of alternative schemes collect and analyze the actual load of every candidate AP. However, these schemes may incur significant latency and signaling overhead in dense wireless local area networks (WLANs). To mitigate such overhead, we propose a user application-based AP selection scheme that considers historical information about AP performance. Without inducing any signaling activity, our scheme monitors the amount of network traffic used by applications and estimates the achievable throughput of APs. Our scheme employs the characteristics of application traffic with the intent of accurately predicting AP performance. Using a measurement study in dense WLAN environments, we show that our scheme achieves higher throughput and lower association latency than those of existing schemes in places highly accessible to the user.
Wireless local area network (WLAN) technology is an important player in providing reliable, portable and high-speed internet connectivity to end users . To provide enough capacity for the large amount of traffic generated by the users, the density of access points (APs) gets much higher than that needed for coverage only . However, high interference levels among APs can cause poor and unpredictable performances . Unfortunately, since most of these APs are under decentralized control, it is complicated to share their resource according to the network condition ; therefore, the users commonly find tens of APs in each scan, which vary significantly with regard to the quality of internet connection .
The proposed scheme is related to two research areas: AP selection and application traffic classification.
We conducted measurement studies to explore the characteristics of dense WLANs and user application traffic and to analyze the effect of user mobility on WLAN connectivity.
On the basis of the observations in the previous section, we propose a UAAS scheme that considers the historical information about AP performance. The goals of our scheme are as follows. First, it accurately examines the achievable throughput of each accessible AP which will be used as the historical information. Second, our scheme mitigates the latency and signaling overheads for obtaining the historical information in dense WLANs.
In this section, we evaluate the performance of the proposed AP association scheme. We utilized our ground-truth measurements to simulate the following two scenarios:
We found through the measurement studies that a user often detects tens of APs and there is a significant range in their achievable throughput in dense WLANs. However, the AP selection schemes based on received signal strength result in poor user experience in many situations. Other alternative schemes incur additional latency and signaling overhead in probing every single candidate AP in dense WLANs. To address this problem, UAAS monitors the amount of network traffic used by applications while the AP is being used. When the user re-encounters the AP, the historical application traffic is used to predict its achievable throughput. Our scheme improves the prediction accuracy by considering the characteristics of each type of application traffic. The results of our experiments using the measurements collected in actual dense WLAN environments and from actual mobile users showed that UAAS outperformed the previous schemes with respect to the achievable throughput and the association latency in the places highly accessible to the user. In addition, the simulation using NS-3 showed that our probabilistic and delayed switching policies achieved higher system throughput and fewer reassociations compared to the deterministic and immediate switching policies.