This article describes the main features of Multiprotocol Label Switching (MPLS), a standard architecture proposed by the IETF that integrates label swapping forwarding with network layer routing. The role of MPLS in overcoming...
moreThis article describes the main features of Multiprotocol Label Switching (MPLS), a standard architecture proposed by the IETF that integrates label swapping forwarding with network layer routing. The role of MPLS in overcoming limitations of overlay models, such as IP over ATM, and of conventional routing in IP networks, is discussed. The main concepts are introduced and the operation of MPLS is explained-classification of packets into Forward Equivalence Classes, label allocation and binding to routes, label distribution, set up of Label Switched Paths and route selection. Finally, support of traffic engineering and Quality of Service mechanisms in MPLS networks is analysed. 1. MPLS-rationale for a new routing and forwarding architecture The success of the Internet is mainly due to its flexible architecture, built upon IP, the ubiquitous internetworking layer protocol. IP networks offer unparalleled scalability and flexibility for the deployment of value-added services and are becoming increasingly attractive for carrying services with hard and soft real-time constraints. This requires extending the traditional best-effort model, designed from the outset for elastic data traffic, with mechanisms that provide differentiated and predictable Quality of Service (QoS) to a wide variety of applications with different requirements [1]. The Internet Engineering Task Force (IETF) has already specified the Integrated Services (IntServ) [2] and the Differentiated Services (DiffServ) [3] models with these goals in mind. Carriers' requirements With the increasing demand and explosive growth of the Internet, service providers require a dependable and controllable network infrastructure that can offer consistent performance. In many cases, the original router-based backbone networks evolved into a two level structure made up of a high speed core network interconnecting edge devices (IP routers) that, in turn, interface with access networks and provide common services to users, such as security, accounting, Virtual Private Networks (VPNs), web hosting, etc. Management of such networks requires powerful traffic engineering techniques, that is, the capability of mapping flows into the physical network topology and evenly distributing traffic over the network links, to achieve an efficient utilization of network resources, avoid congestion and improve network performance.
