CCNP SP · UNIFIED MPLS

Inter-AS Unified MPLS: three ISPs with BGP-LU

Three providers carry one end-to-end LSP between PE1 (ISP1, AS 6001, IOS XE) and PE3 (ISP3, AS 6003, IOS XR), across the transit provider ISP2 (AS 6002, IOS XR). Each ISP runs its own IS-IS and LDP, and nothing of that crosses the borders. Stitch them with BGP labeled unicast: iBGP inside each ISP, eBGP between the border routers. ISP2’s P2 runs no BGP at all; it forwards the traffic only because it is label-switched.

CCNP SP track · Cisco IOS XE + IOS XR · 6 objectives

Open in simulator More CCNP SP labs

Objectives

The simulator ticks each one as soon as the network meets it.

PE1 (6001) and PE3 (6003): peer iBGP with your ASBR between loopbacks, exchange labels, and advertise your Loopback0 /32.
ASBR1 and ASBR3: peer eBGP with ISP2 on the link addresses and exchange labels; set next-hop-self toward your PE. ASBR3 (IOS XR) needs route-policy PASS in and out on its eBGP neighbor.
ASBR2A and ASBR2B (AS 6002, IOS XR): eBGP labeled unicast toward ISP1 and ISP3 with route-policy PASS in and out, and iBGP labeled unicast between your loopbacks with next-hop-self. Remember allocate-label all.
PE1 learns 7.7.7.7/32 from ASBR1 with a label (show bgp ipv4 unicast labels).
PE1 and PE3 have a BGP-LU path to each other’s loopback.
Ping 7.7.7.7 from PE1, sourced from Loopback0: the echo crosses ISP2’s BGP-free P2.

Start this lab

Before you start

Over eBGP the next hop always changes, so every border router allocates its own label for the routes it passes on; the eBGP next hop is directly connected, so only the BGP label crosses the border link. Inside each ISP the BGP next hop is a loopback reached over LDP. IOS XR accepts and sends no eBGP routes without a route-policy in each direction, and labels nothing without allocate-label all. On real IOS XR an eBGP labeled-unicast neighbor also needs a /32 static route to its address; RouteCraft does not ask for it.

Concept

Inter-AS Unified MPLS: BGP-LU between providers

One LSP can cross several autonomous systems. Inside each provider, iBGP labeled unicast carries the PE loopbacks to the border routers, and LDP carries the packets between BGP next hops. Between providers, the border routers (ASBRs) peer over eBGP labeled unicast on their shared link. The next hop changes on every eBGP hop, so each ASBR allocates its own label. This is the idea behind inter-AS option C.

Concept

Unified MPLS: BGP labeled unicast across IGP domains

Large networks split their IGP into domains (IS-IS areas or separate instances) so each one stays small. LDP then only builds paths inside a domain. A Level-1 PE knows just a default route, and LDP gives no label for a default route. Unified (seamless) MPLS stitches the domains with BGP labeled unicast: each PE advertises its loopback with a label, and the border routers reflect it with next-hop-self and a label of their own.