Multi-User Resource Allocation and Rate Adaptation for IEEE 802.11 WLANs

The Hong Kong University of Science and Technology
Department of Computer Science and Engineering


PhD Thesis Defence


Title: "Multi-User Resource Allocation and Rate Adaptation for IEEE 802.11 
WLANs"

By

Mr. Arafet BEN MAKHLOUF


Abstract

Wireless Local Area Networks (WLANs) are becoming increasingly popular due to 
the recent availability of affordable devices providing multiple and high rate 
capabilities. New PHY and MAC layer enhancements have been introduced in 
different IEEE 802.11 WLAN standards. These improvements have given birth to a 
wide range of bandwidth channels (up to 160MHz) and very high transmission data 
rates (> 1Gbps) to keep up with current and upcoming multimedia applications. 
Optimizing the performance of different wireless networks for emerging network 
applications is an important and highly challenging issue.

QoS is very important for high throughput WLANs. Scheduling the access of flows 
in a non-interfering manner to avoid packet loss and adapting rates to changing 
channel conditions are important aspects of QoS in wireless networks. The goal 
of the rate adaptation scheme is to select the best set of transmission 
parameters for a given user considering its channel quality. Similarly, 
sub-channel scheduling access can be seen as selecting the best set of users to 
allocate subcarriers according to their channel state information. As data 
traffic has diverse characteristics and different transmission requests, an 
advanced version of Multimedia-MAC should exist to dynamically adjust the 
bandwidth allocation according to the instant traffic load status so that both 
network channel utilization and QoS for different types of traffic streams are 
dramatically enhanced.

This thesis makes two main contributions. The first is a design a new practical 
rate control algorithm for 802.11n WLANs, based on a probing system that 
guarantees that it is has Long-Term Stability and Short-Term Responsiveness 
(L3S). We then implement it in commercial devices using the Ath9k driver 
without modifications to the existing standard. The new rate adaptation 
classifies transient and sustained changes in the link conditions. Then, it 
controls both short-term and long-term channel quality variations respectively 
by continuously monitoring the transmission history and intelligently probing 
at new data rates that may outperform the current one. Our proposed rate 
control algorithm adapts rapidly to these changes by adjusting the efficient 
transmission rate. Thus, it optimizes the throughput (or delay) performance on 
a wireless link. The second part consequently deals with the major reasons for 
the inefficiency of different WLANs. The current MAC layer randomly allocates 
the entire channel to only one user as a single resource. Indeed, one way to 
significantly improve WLAN performance is to effectively reduce the channel 
width and create different sized sub-channels. Based on the user's channel 
conditions and QoS requirements, the Physical layer resources can be 
dynamically allocated to several users at the same time. Thus, we present a 
novel sub-channels access approach, titled MU-Access, with new features 
suitable for the forthcoming high-speed MIMO-based WLAN products. We shall 
evaluate its performance compared to state-of-the-art systems, using NS-3 under 
a variety of network conditions. Our experiments then demonstrate that our 
scheme does indeed enhance IEEE 802.11 device performance with a much higher 
degree of throughput. Also, we propose to profit from the sub-channels access 
approach and adapt the rate separately for each block of subcarriers according 
to the instantaneous channel conditions, for improvement.


Date:			Friday, 22 August 2014

Time:			2:00pm - 4:00pm

Venue:			Room 3501
 			Lifts 25/26

Chairman:		Prof. Amine Bermak (ECE)

Committee Members:	Prof. Mounir Hamdi (Supervisor)
 			Prof. Jogesh Muppala (Supervisor)
 			Prof. Brahim Bensaou
 			Prof. Gary Chan
 			Prof. Khaled Ben Letaief (ECE)
 			Prof. Chengshan Xiao (Elec. & Comp. Engg.,
 					      Missouri U. of Sci. & Tech.)


**** ALL are Welcome ****