Fair medium access control scheme for packet ring networks

In recent years, the rapidly increasing volume of data traffic in Metropolitan Area Networks (MANs) is challenging the capacity limits of existing transport infrastructures based on circuit-oriented technologies such as Synchronous Optical Network (SONET). Packet ring networks have attracted wide attention as the transport technology for data services in MAN. It combines the resilience and deterministic performance of SONET and the data friendliness of Ethernet switches. However, since packet rings allow concurrent transmission of packets over different segments, the ring bandwidth becomes a shared medium. Hence, a key challenge for packet rings is to design a Medium Access Control (MAC) scheme that ensures all nodes have fair access to the ring.; In this study, we investigate the existing MAC schemes including the IEEE RPR standard and its enhancements, in which ring access is provided through fair rate estimation and feedback control. The existing approaches suffer from some performance limitations such as large access delay and throughput loss.; The deficiencies of the existing schemes motivate the research of this thesis: fair MAC scheme for packet rings. Our solution is the design of a novel MAC scheme, called Virtual Source Queuing (VSQ). VSQ offers guaranteed fair medium access by providing individual virtual queues and fair queuing to all the competing sources. We also propose a simple flow control scheme called One-Hop-Backpressure (OHB) to provide lossless transit paths for VSQ. OHB exploits the source information provided by VSQ to effectively control the transit buffer occupancy through implicitly controlling the input rate of the traffic sources. As a result, VSQ requires less transit buffer size to achieve a lossless transit path than RPR does.; We conduct in-depth performance evaluations of VSQ and RPR. Our analysis and simulation results reveal that VSQ not only meets all the MAC design goals for packet rings, such as guaranteed fairness and maximum throughput, but also offers faster fairness convergence, less access delay and less transit buffer occupancy in comparison to RPR.