| Enterprises of all sizes are looking for new solutions to meet the ever changing demands on data communication as traditional data services are expensive to scale, limited and inflexible in their service options, operationally complex, and painfully slow to upgrade. This is especially true in Metro area networks where service providers are falling far short of enterprise requirements for bandwidth with flexibility and scalability.; The flexibility, scalability, simplicity and low cost of Ethernet technology make it an ideal networking technology for Metro networks. Ethernet dominates the current LAN realizations and efforts are underway to make Ethernet as an end-to-end technology spanning across LAN and MAN (Metro Area Network). Deploying Ethernet in the Metro domain will require several upgrades including end-to-end QoS guarantees, protection mechanisms and service performance monitoring. Further, the new solution(s) must support the performance requirements of a company's growing network and at the same time preserve their investment in network infrastructures.; IEEE has several supporting standards to the basic Ethernet standard for extending its application to the metro area. Even though these standards suggest methods for improving QoS, they are inadequate to escalate Ethernet to a carrier grade technology in the metro domain. To that end, there are some recent works specifically addressing the QoS issues of Ethernet for larger networks. Some results are promising however, far from comprehensive.; In response, our research focuses on core QoS enabling mechanisms such as bandwidth accounting, traffic engineering, and fast restoration to provide a holistic solution for QoS. First, we address the problem of traffic engineering in Ethernet which is one of the integral components of QoS provisioning. We propose a scheme based on the generation and management of multiple spanning trees for a near optimal traffic distribution. Second, we propose an algorithmic approach for constructing multiple spanning tree regions in the enterprise network domain which will provide better recovery time, reusability of VLAN tags, protection from failures, and optimal broadcast domain size. Survivability issues in these environments become more important as network links carry large amounts of traffic. We develop a distributed algorithm to address network failures, based on a fast failure recovery spanning tree (FFRST) scheme, which restores lost facilities within tens of milliseconds. Recovery algorithm is localized around the point of failure on the spanning tree, thus avoiding disruption of the entire network. Further, in this section we discuss possible enhancement to our recovery method. Despite the advances we make in the dissertation, several issues and challenges still exist and we provide some pointers towards that as future work. |
