Date of Award

2015

Document Type

Thesis

Publisher

Edith Cowan University

Degree Name

Doctor of Philosophy

School

School of Engineering

Faculty

Faculty of Health, Engineering and Science

First Supervisor

Professor Daryoush Habibi

Second Supervisor

Dr Iftekhar Ahmad

Abstract

Exponential growth in the volume of wireless data, boosted by the growing popularity of mobile devices such as smart phones and tablets, is forcing telecommunication industries to rethink network design, and focus on developing high capacity mobile broadband networks. Accordingly, researchers have undertaken developmental work for an integrated wireless-optical broadband access network (WOBAN). Passive optical networks (PONs) and fourth generation (4G) wireless networks are two major candidate technologies for the WOBAN. PON is a wired access technology, well-known for its high capacity, whereas 4G is a wireless broadband access technology, popular for its ease of deployment and ability to offer mobility. Integration of PON and 4G technologies, as a wireless-optical broadband access network, offers advantages such as extension of networks in rural areas, support for mobile broadband services, and rapid deployment of broadband networks. However, these two technologies have different design architectures for handling broadband services which require Quality of Service (QoS), for example, 4G networks use traffic classification for supporting different QoS demands whereas PON does not differentiate between traffic types. This integrated network must also be energy efficient, as a green broadband access network, without hindering QoS. While these technologies both use sleep mode, they differ in their power saving mechanisms.

This thesis first addresses a QoS solution for the incompatibility between these technologies. Service class mapping is proposed in Chapter 3 for the integrated WOBAN, based on the M/G/1 queuing model supported by an innovative priority scheduler. Once class mapping is deployed, a power saving mechanism can be devised by exploiting traffic differentiation. Specifically, a class-based strategy is proposed which helps optimise the sleep period for the terminal units of the optical network, without compromising QoS.

Since the optical network involves control and terminal nodes, both of which consume power, this thesis proposes an energy efficient mechanism that involves both components. In contrast, other published strategies (Chapter 2) have only considered the terminal units. Chapter 4 presents the mechanism for enabling global sleep (control and terminal nodes) and local sleep (terminal nodes), based on the available traffic's class structure. This mechanism enables sleep for different components within the bandwidth allocation by adapting the switching between predefined polling cycle lengths.

As the WOBAN is comprised of both wireless and optical parts, a dynamic resource management mechanism is needed which responds to changing daily traffic patterns across a green integrated network. Consequently, Chapter 5 proposes a mechanism which dynamically adapts the polling cycles, of the optical and wireless parts of the network, to the changing traffic volume and class composition. Tailored sleep durations for the components of the WOBAN are facilitated within the resource management regime, as these components differ in their ability to function efficiently if management of the sleep periods is not responsive to the changing traffic volumes and class composition.

This dissertation creates new knowledge by seamlessly integrating the two parts of WOBAN and introducing differentiated, class-based sleep for the components of the hybrid network to help realise a green WOBAN.

Share

 
COinS