Reference:
http://www.cisco.com/c/en/us/support/docs/multiprotocol-label-switching-mpls/mpls/118846-config-mpls-00.html#anc2
Why Unified MPLS?
- Greater scalability
- Faster convergence
- Smaller IGP/LDP domains, no flat IGP/LDP domain as with traditional MPLS.
Cisco Unified MPLS uses BGP-4 with label information exchange (RFC3107). When BGP distributes a route, it can also distribute an MPLS label that is mapped to that route. The MPLS label mapping information for the route is carried in the BGP update message that contains the information about the route. If the next hop is not changed, the label is preserved and the label changes if the next hop changes. In Unified MPLS, the next hop changes at Area Border Routers (ABRs).
When you enable RFC 3107 on both BGP routers, the routers advertise to each other that they can then send MPLS labels with the routes. If the routers successfully negotiate their ability to send MPLS labels, the routers add MPLS labels to all outgoing BGP updates.
The label exchange is needed in order to keep the end-to-end path information between segments. As a result, each segment becomes small enough to be managed by operators and at the same time there is circuit information distributed for path awareness between two different IP speakers.
Traditional MPLS uses two labels - Service Label and LDP/IGP label
Along with the traditional MPLS labels, unified MPLS adds one additional label which is BGP3107 label.
How the packet forwarding is done in Unified MPLS?
- On PE11, Prefix A is known via BGP session with PE31 as next-hop PE31 and PE31 is recursively reachable via P1 with BGP label 100. PE11 received IPv4 + Label information from P1 as BGP updates because it is enabled with the RFC 3107 feature in order to send the IPv4 + Label information.
- P1 is reachable from PE11 via intradomain LDP LSP and it adds another LDP label on top of the BGP label. Finally, the packet goes out of the PE11 node with three labels. For example, the 30 L3VPN service label, the 100 BGP label, and the 200 LDP IGP label.
- The LDP top label continues to swap in intradomain LDP LSP and the packet reaches P1 with two labels after Penultimate Hop Popping (PHP).
- P1 is configured as inline Route Reflector (RR) with next-hop self and it joins two IGP domains or LDP LSP.
- On P1, the next hop for PE31 is changed to P2 and the update is received via BGP with IPv4 + Label (RFC3107). The BGP label is swapped with new label because next-hop is changed and the IGP label is pushed on top.
- The packet goes out of the P1 node with three labels and service label 30 is untouched. That is, the 30 L3VPN service label, 101 BGP label, and 201 LDP label.
Note the change in BGP label in this step.
- The LDP top label swaps in intradomain LDP LSP and the packet reaches P2 with two labels after PHP.
- On P2, the next hop for PE31 is changed again and it is reachable via IGP. The BGP label is removed as an implicit-null BGP label is received from PE31 for PHP.
- The packet leaves with two labels. For example, the 30 L3VPN service label and the 110 LDP label.
- On PE31, the packet arrives with one label after PHP of the LDP label and based on the service label 30. The unlabeled packet is forwarded to the CE31 destination under Virtual Routing and Forwarding (VRF).
Configuration commands related to Unified MPLS is covered in this link on cisco.com
http://www.cisco.com/c/en/us/support/docs/multiprotocol-label-switching-mpls/mpls/116127-configure-technology-00.html#anc2
