Admission Control and Bandwidth Allocation for Class A Traffic in RPR Networks
Zhenning Xu
Abstract
Zhenning Xu
Abstract
In this thesis, we study the admission control and bandwidth allocation methods for classA traffic in RPR networks. First, we investigate the performance of classA traffic under the current RPR protocol. The simulation results show that RPR networks can support low-delay classA traffic even if the networks are congested with classB and classC traffic. The low-delay performance, however, is subject to the condition that the load of classA traffic must be properly controlled. Consequently, an admission control mechanism must be used for classA traffic. In this thesis, several admission control algorithms are studied. They are the Simple Sum algorithm, the Measured Sum algorithm, and the Equivalent Bandwidth algorithm. The simulation results show that the Equivalent Bandwidth algorithm is the most suitable to use as the admission control mechanism for classA traffic. The admission control mechanism makes admission decision based on the available bandwidth allocated to the classA traffic. The existing RPR standard assumes the bandwidth allocated for classA traffic at each node is fixed. The fixed bandwidth allocation introduces inflexibility and inefficient use of bandwidth for classA traffic. In this thesis, three bandwidth allocation algorithms are proposed to dynamically allocate bandwidth for classA traffic. These algorithms have different levels of complexity and can be applied to different traffic environments. Simulation results show that the proposed algorithms improve the bandwidth efficiency of the RPR networks. The proposed algorithms are also readily integrated with the existing Internet Quality of Services (QoS) paradigms such as Diffserv and RSVP services.
A significance statement is not available in the OpenAlex record.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
In this thesis, we study the admission control and bandwidth allocation methods for classA traffic in RPR networks. First, we investigate the performance of classA traffic under the current RPR protocol. The simulation results show that RPR networks can support low-delay classA traffic even if the networks are congested with classB and classC traffic. The low-delay performance, however, is subject to the condition that the load of classA traffic must be properly controlled. Consequently, an admission control mechanism must be used for classA traffic. In this thesis, several admission control algorithms are studied. They are the Simple Sum algorithm, the Measured Sum algorithm, and the Equivalent Bandwidth algorithm. The simulation results show that the Equivalent Bandwidth algorithm is the most suitable to use as the admission control mechanism for classA traffic. The admission control mechanism makes admission decision based on the available bandwidth allocated to the classA traffic. The existing RPR standard assumes the bandwidth allocated for classA traffic at each node is fixed. The fixed bandwidth allocation introduces inflexibility and inefficient use of bandwidth for classA traffic. In this thesis, three bandwidth allocation algorithms are proposed to dynamically allocate bandwidth for classA traffic. These algorithms have different levels of complexity and can be applied to different traffic environments. Simulation results show that the proposed algorithms improve the bandwidth efficiency of the RPR networks. The proposed algorithms are also readily integrated with the existing Internet Quality of Services (QoS) paradigms such as Diffserv and RSVP services.
Key concepts: Dynamic bandwidth allocation, Computer science, Bandwidth allocation, Admission control, Bandwidth (computing), Network traffic control, Quality of service, Computer network