On the Performance of MPLS TE Queues for QoS Routing
暂无分享,去创建一个
Traffic engineering (TE) refers to techniques and processes to route traffic through a network on a path other than that would have been chosen if standard routing methods had been used. The goal of traffic engineering to a service provider is to maximize the utilization of network resources, and/or enhance QoS a service provider can offer. To justify the increase in network operational complexity associated with traffic engineering, TE must enable new service offerings, reduce the overall cost of operations, maximize potential revenues and increase customer satisfaction. In a large network, it is possible that available network bandwidth is not efficiently utilized because the intradomain routing protocol, such as OSPF, finds path based on a single “least-cost” scalar metric for each destination. This least cost route may not have enough resources to carry all the traffic, or satisfy all the SLA (service level agreement) requirements of passing traffic. Congestion, either for aggregate traffic or for per class traffic, at certain hot spots, can result in sub-optimal use of network resource. Multi-Protocol Label Switching (MPLS) is an advanced forwarding scheme which extend routing with respect to packet forwarding and path controlling. MPLS TE can relieve congestion for aggregate traffic. In addition, when QoS is deployed in a service provider’s network, TE can also help to relieve per class congestion and provide better QoS guarantees. MPLS TE provides a technique, more elegant and efficient than IP source routing, to allow traffic travel down a path different from conventional IGP destination based hop-by-hop routing. The path is pre-determined at tunnel setup time. Routers along the path do not have to examine the IP header of every passing packet. The basic idea of MPLS involves assigning short fixed length labels to packets inside an MPLS cloud. Throughout the MPLS domains, the labels attached to packets are used to make forwarding decisions. It allows decoupling of the information used for forwarding (a label) from the information carried in the IP header. MPLS TE, using RSVP signaling mechanism [1], injects the notion of a connection to connectionless IP through nailed-up label switched paths (LSP). MPLS TE provides capabilities to specify explicit path for the LSP (ER-LSP) before the LSP is established. We’ll refer the nailed-up LSP as MPLS TE tunnel, or simply tunnel, in this paper. The tunnel explicit routing capability allows routing flexibility. It allows paths, with unequal OSPF cost, to share traffic load. In addition, the Fast Reroute (FRR) feature in MPLS TE allows path restoration within 100 ms in case of link or node failure. In this paper we propose a MPLS TE tunnel mechanism for packet forwarding, which can guarantee the service of real time applications such as VoIP and videoconferencing.
[1] Shivendra S. Panwar,et al. The effective bandwidth of Markov modulated fluid process sources with a generalized processor sharing server , 2001, GLOBECOM'01. IEEE Global Telecommunications Conference (Cat. No.01CH37270).