Date of Award
Doctor of Philosophy (PhD)
Physical constraints, bandwidth constraints and host mobility all contribute to the difficulty of providing Quality of Service (QoS) guarantees in wireless networks. There is a growing demand for wireless networks to support all the services that are available on wired networks. These diverse services, such as email, instant messaging, web browsing, video conferencing, telephony and paging all place different demands on the network, making QoS provisioning for wireless networks that carry multiple classes of traffic a complex problem. We have developed a set of admission control and resource reservation schemes for QoS provisioning in multi-class wireless networks.
We present three variations of a novel resource borrowing scheme for cellular networks that exploits the ability of some multimedia applications to adapt to transient fluctuations in the supplied resources. The first of the schemes is shown to be proportionally fair: the second scheme is max-min fair. The third scheme for cellular networks uses knowledge about the relationship between streams that together comprise a multimedia session in order to further improve performance. We also present a predictive resource reservation scheme for LEO satellite networks that exploits the regularity of the movement patterns of mobile hosts in LEO satellite networks. We have developed the cellular network simulator (CNS) for evaluating call-level QoS provisioning schemes. QoS at the call-level is concerned with call blocking probability (CBP), call dropping probability (CDP), and supplied bandwidth. We introduce two novel QoS parameters that relate to supplied bandwidth—the average percent of desired bandwidth supplied (DBS), and the percent of time spent operating at the desired bandwidth level (DBT).
Rizvi, Mona E..
"QoS Provisioning for Multi-Class Traffic in Wireless Networks"
(2004). Doctor of Philosophy (PhD), Dissertation, Computer Science, Old Dominion University, DOI: 10.25777/qtv4-et29