What are the major drawbacks of traditional IP forwarding and how does MPLS solve them?

Speed - the forwarding paradigm employed by MPLS is based on the notion of label swapping. The routers can forward packets faster, since they only need to inspect a short label, instead of the traditional  longest-match forwarding used at the network layer.

Connectionless Service, and Connectionless VPN's - unlike FR and ATM, a significant technical advantage of MPLS VPNs is that they are connectionless. TCP/IP is connectionless. This means that no prior action is necessary to establish communication between hosts, making it easy for two parties to communicate. Legacy VPN's using connectionless TCP/IP networks,  had to add a connection-oriented, point-to-point overlay on the network.  Even though they ran on a connectionless network, the traditional VPN's could not take advantage of the ease of connectivity, because for security purposes - they had to constrain the network through the use of tunnels and encryption.  But MPLS offer a connectionless VPN, and therefore you do not need the tunnels and encryption for network privacy, thus eliminating significant complexity.

Any-to-Any Connectivity - an MPLS-based IP VPN solves this problem because it creates a private, meshed network of the customer’s locations. An MPLS-based IP VPN solves the proverbial “N squared”7 scaling problem of Layer2 networks such as Frame Relay and ATM where many dedicated permanent virtual circuits (PVCs)/Virtual Circuits(VCs) are required if all customer locations must communicate with each other.

With PVC-based network connectivity (Frame Relay, ATM, etc), the number of links required for a full mesh is (N*(N-1))/2.  For a4-node network, you need 4(4-1) /2 =6 PVC’s.  But for a network of500 nodes, you need 500(500-1) /2 =124,750 PVC’s !!!  But with an500 node MPLS network, you need zero PVC’s.

With an MPLS-based IP VPN, authorized customer locations are provided with VPN connectivity to all other authorized customer sites within the same VPN in a completely meshed fashion, similar to a private Internet in operation. But unlike the Internet, an MPLS-based IP VPN is as secure as frame relay or ATM, and is designed to prevent traffic from unauthorized sources from accessing a customer’s VPN.

Smooth Layer2-to-Layer3 Transition - support for Layer2 protocols also enables customers to smoothly transition from a traditional Layer2 network, such as frame relay and ATM, to an MPLS-based IP VPN without making expensive and disruptive changes in their equipment or addressing. For carriers, Layer2 support provides a platform for network convergence as it enables legacy Layer2 services to be transported across an MPLS cloud.

Flexible Addressing Support -  an MPLS-based IP VPN supports any type of customer IP addressing, thereby providing maximum flexibility to enterprises – renumbering is not required when connecting to a carrier’s MPLS network. Precisely because of the privacy of the VPN; the customer’s IP addresses are only visible within that VPN and not beyond. There is no need to renumber internal networks.

Each customer’s VPN is separated logically from every other VPN.  So the network is very forgiving of IP addresses, so long as each IP address in a given VPN is unique.  This allows reuse of the same addresses, so long as they are in separate VPN’s.

NAT (Network Address Translation) is not needed – NAT is a major headache for large networks, because every private address is translated to a public address.  It requires extra bookkeeping and router processing  .  .  . and it does not work well with IPsec, since authentication of packets will fail if the address has been tampered with.  But with MPLS, the private addresses are perfectly legal !!  So there is no need to NAT them – let them stay as they are, and turn off NAT in the customer’s routers.

Routing Simplicity - connecting to an MPLS-based IP VPN service is simple for an enterprise customer. Customers need not run MPLS within their networks in order to take advantage of an MPLS-based IP VPN; MPLS is managed completely by the carrier, giving enterprises the benefits of a secure, scalable, virtual private network without administrative complexity.  Either static routing or BGP48 routing protocols are supported between the carrier’s network equipment (called the Provider Edge or PE, in MPLS terminology) and the customer’s router (called the Customer Edge or CE).