mpls traffic-engineering options

can somebody explain different b/w mpls traffic-engineering options?
bgp-igp-both-ribs vs mpls-forwarding

root@R2# set traffic-engineering ?
Possible completions:
bgp BGP destinations only
bgp-igp BGP and IGP destinations
bgp-igp-both-ribs BGP and IGP destinations with routes in both routing tables
mpls-forwarding Use MPLS routes for forwarding, not routing
[edit protocols mpls]

both retain routes in inet.3 table and also copy inet.3 routes into inet.0

Find more posts tagged with

Comments

Aldur

The one major difference is that 'mpls-forwarding' allows you to use the non-mpls related routes for routing and the mpls based routes for forwarding.

So lets look at what happens when you use 'bgp-igp-both-ribs'. The routes from the inet.3 table get moved into the inet.0 table. Those routes, LDP or RSVP, have a much lower preference then any other routes in that RIB so they are chosen as the active routes. This can be a problem since routing policies can only act upon active routes.

By using the 'mpls-forwarding' command instead it allows you to use the mpls based routes for forwarding but keep the BGP/IGP routes active for routing purposes, such as routing policies.

So in production networks I mainly see 'mpls-forwarding' used for this reason, it really gives the best of both worlds.

densma

Thanks you!!

Aldur wrote: »

The one major difference is that 'mpls-forwarding' allows you to use the non-mpls related routes for routing and the mpls based routes for forwarding.

So lets look at what happens when you use 'bgp-igp-both-ribs'. The routes from the inet.3 table get moved into the inet.0 table. Those routes, LDP or RSVP, have a much lower preference then any other routes in that RIB so they are chosen as the active routes. This can be a problem since routing policies can only act upon active routes.

By using the 'mpls-forwarding' command instead it allows you to use the mpls based routes for forwarding but keep the BGP/IGP routes active for routing purposes, such as routing policies.

So in production networks I mainly see 'mpls-forwarding' used for this reason, it really gives the best of both worlds.

dwater2010

Aldur so bgp-igp option will make BGP to establish over MPLS Tunnel path?

where as in other only forwarding traffic will use MPLS tunnel and for control protocol it use inet.0 table only?

Aldur

dwater2010 wrote: »

Aldur so bgp-igp option will make BGP to establish over MPLS Tunnel path?

Well, yes, but I think you're missing the point here. If you have two BGP neighbors and peering addresses are found in the inet.0 routing table as an LDP route then yes the packets will tranverse the LSP to get to the peer. But this is true of bgp-igp and mpls-forwarding.

Try this example for understanding. You have a router that is running BGP and it has a BGP route of 10.10.10.1/32 that it needs to send to another BGP speaker. If the bgp-igp statement is used and your router also has an LSP to 10.10.10.1 then the MPLS route for 10.10.10.1 becomes active in the inet.0 routing table. This will stop your router from advertising the 10.10.10.1 route to another BGP router.

On the other hand, if you have used mpls-forwarding instead the 10.10.10.1/32 BGP route will be passed on to the other BGP speaker and the LSP will still be used to forward traffic to the 10.10.10.1 destination.

Make sense?