ABSTRACT
Initially, the objective for developing the Internet was purely to transport data from one terminal to the other without considerations for QoS management (best effort was the aim), speed, scalability, and traffic engineering. As time progressed, the need for QoS increased, leading to protocols such as MPLS (MultiProtocol Label Switching). The original driver for introducing the MPLS was to make routers faster. It was observed that ATM switches, which used virtual path switching, were faster than IP routing. Trying to match a long IP address in a router is slower than using purely short-length labels lookup in routers. Achieving this allows devices to match the performance of ATM switches while doing the same job of data routing. Furthermore, MPLS enables integration of IP and ATM, so that ATM can be mapped to IP. In conventional IP packet routing, the destination IP address in the packet header is used to forward a packet to the destination. MPLS helps to decouple this IP packet-forwarding dependence to the information carried in the IP packet header. By not first inspecting the header of the packet before it decides to send a packet, MPLS proves to work independent of the source of data. It leads to the emergence of new routing functions and forwarding paradigms not supported by IP routing. For example, MPLS enables multiprotocol lambda switching, virtual private networks, layer 2 transport, traffic engineering, and guaranteed bandwidth services.
