| The goal of this project is to investigate various QoS mechanisms relevant for Ethernet, improve where necessary and then quantify their interaction and influence on end-to-end performance. We aim at assessing the influence of various schemes and parameters, not just at the Ethernet level but also on (higher) application-layer performance. Thus we address both inter-layer and intra-layer issues. By applying or developing new modelling techniques, we aim at obtaining generic results and guidelines quantifying the influence of network and traffic parameters on the QoS performance. This will assist the optimal deployment of Ethernet services in MENs. Given the set time frame it is not possible to study and investigate all mechanisms. The mechanisms, which are not the main research focus, are packet marking, traffic classification and traffic shaping. Packet marking and traffic classification have a rather straightforward approach and are restricted by the number of bits available in the packet header. Further information on these possibilities is available in the IEEE 802.1ad working group [14]. Traffic shaping is a topic extensively researched in literature. Traffic shaping automatically occurs at every queue. Since the enhancements in literature require expensive buffer space and the mechanism does not face Ethernet specific issues we do not include it as a main point of interest in this proposal. It is important to note that we will take these mechanisms into account for the end-to-end performance analysis but we will not research them individually. The QoS mechanisms of our focus in this proposal will be o Traffic policing o Queuing/Scheduling o Buffer management o Congestion avoidance o Interaction of QoS mechanisms and resulting multi-layer performance |
