Quality of service and performance optimisation in wireless local area networks

Wireless local area network (WLAN) technology has experienced tremendous growth over the past decade, primarily attributed to the explosive growth of the Internet and ever increasing user demands for feature-rich multimedia content and services in combination with its ease of deployment and relatively inexpensive infrastructure costs. With increases in physical layer (PHY) data rates from 1-2Mb/s in the original IEEE 802.11 specification to an unprecedented 600Mb/s in upcoming standards (with > 1Gb/s being considered for future-generation systems), 802.11 technology has truly become one of the most remarkable standardisation achievements in the telecommunications arena today. In addition to increased PHY data rates, 802.11 has evolved from its earlier incarnations to become a ubiquitous, high-throughput wireless access technology capable of advanced quality of service (QoS) provisioning. However, commercial implementations often provide only “besteffort” QoS support, are implemented with partial or manufacturer specific QoS enhancements, or use sub-optimal medium access control (MAC) reference specifications with limited ability to provide optimal system throughput and QoS provisioning. Furthermore, the dynamic adaptation of MAC parameters for achieving optimal system performance is out of scope of the current standard specifications, and thus provides an open and challenging research area. This thesis investigates advanced MAC techniques designed to maximise the efficiency and system utilisation of the 802.11 protocol, through the application of novel overhead reduction and resource management techniques. Achieving optimal system performance whilst meeting heterogeneous QoS requirements is highly challenging. Recognising this opportunity, an analysis of optimal automatic repeat-request (ARQ) modes for various traffic classes and channel conditions is presented, with a novel adaptive ARQ scheme developed. Further, a distributed adaptive block size algorithm is presented; designed to improve throughput, delay and packet loss protection through the dynamic adaptation of the Block Ack ARQ mode block size parameter, based on link state metrics and QoS requirements. Finally, a novel scheme for reducing MAC protocol overhead by altering the underlying protocol timing constraints is presented. The proposed techniques address a number of performance enhancement opportunities that have received little attention thus far. Evaluation is performed where possible using a theoretical framework, with comprehensive simulation studies used to gain a detailed understanding of the performance benefits when operating under more realistic scenarios. Lastly, whilst targeted towards application in state of the art 802.11 systems, in general, the techniques proposed in this thesis remain relevant to other wireless MAC protocols where ARQ techniques are employed.