Hi Laurent,

Many thanks

What about this?

http://www.cisco.com/en/US/docs/ios/mpls/configuration/guide/mp_te_frr_node_prot.html#wp1477142

Actually i tried to lab FRR using RSVP hellos on a c7200 running 12.4(20)T and the tunnel is not coming up.

VB4990-CPE-1-1#sh mpls traffic-eng tunnels brief

Signalling Summary:

LSP Tunnels Process: running

Passive LSP Listener: running

RSVP Process: running

Forwarding: enabled

Periodic reoptimization: every 3600 seconds, next in 1619 seconds

Periodic FRR Promotion: Not Running

Periodic auto-bw collection: every 300 seconds, next in 119 seconds

TUNNEL NAME DESTINATION UP IF DOWN IF STATE/PROT

VB4990-CPE-1-1_t0 10.100.240.3 - unknown up/down

VB4990-CPE-1-1_t100 10.100.240.3 - unknown up/down

Displayed 2 (of 2) heads, 0 (of 0) midpoints, 0 (of 0) tails

VB4990-CPE-1-1#

Do you know if there is any sample config to use for MPLS TE FRR tunnel protection available?

Giulio,

This is MPLS fast reroute. The original post was referring to IP fast reroute.

As for the issue you run into is concerned, here is a couple of things you should had to the configuration for the tunnel to come up:

1. mpls traffic-eng tunnels on all the physical interface used to setup the tunnel (fa1/0 and fa1/1 in the config you provided)

2. For rsvp hello to work, you need to configure "ip rsvp signalling hello" on fa1/0 and fa1/1.

Regards

Thanks a lot, its all ok now.

Hi,

I have heard that IP FRR and in latest IOS-XR MPLS FRR ( per link LFA) features with BGP PIC EDGE and CORe can give SDH kind of convergance and one don't need ot use normal TE and FRR.

This all will be avilable going forward in IOS-XR 12k/CRS etc.

Is that correct ?

Hi,

IP FRR will allow you to have same convergence time as MPLS-TE FRR but on an IP only backbone.

BGP PIC allows you to converge at the IGP speed regardless of the number of BGP prefixes.

BGP PIC Core is already available on IOS-XR and IOS (7600 and GSR both with some restrictions).

BGP PIC Edge will be available in IOS-XR and IOS for 7600.

HTH

Laurent.

Do you know when will MPLS FRR ( i.e. LDP FRR) will be avilable for MPLS backbone ? I know it is only going to avilable in IOS-XR first but any idea on release ?

Regards,

Chintan

Chintan,

For such question, the best entry point is your Cisco account team.

Laurent.

Hi Laurent,

From latest 2011 Cisco Networkers, I got to know that IP FRR with LDP support for say IS-IS is now avilable in 4.0.1 for both per-link and per-prefix.

Is that correct ? if so, Do i need to configure any specfic under LDP or it is with in software once you enable ip frr under IS-IS.

Also if you can direct to any good CCO link to verify show commands for IP FRR to see which is backup path etc .

Thanks,

Chintan

http://www.cisco.com/en/US/docs/routers/crs/software/crs_r4.0/mpls/configuration/guide/gc40crs1book_chapter1.html#reference_063CBD50AC624F28B69D6B2173B53A75

HTH

Laurent.

Hi Laurent,

I read that 3.9.x onwards ther eis BGP PIC Edge support with unipath and before that it was must to use multipath.

What is difference between unipath vs multipath  ? Do I get any advantage of memory utilization etc ?

Regards,

Chintan

Hi,

Multipath means you selected two BGP path to do load-balancing so both path are programmed in hardware by default so it's very easy to achieve BGP PIC Edge.

Unipath means you selected only the Best path. By default only this one is programmed in hardware so you can't achieve BGP PIC Edge in this case. The feature has been enhanced so a backup path is selected and also programmed in addition to the Best one.

HTH

Laurent.

Thanks for clarification Laurent.

So What I understand is this new enhancment in 3.9.x will keep now backup path also in FIB and will be programmed in HW.

And here we can have primary/backup link concept if customer wish to have and still provide better convergance should primaly link/PE fails....

If we use multipath then traffic will be loadbalanced than having primary/backup link but we can achieve....

Now When it comes to BGP PIC EDGE , I guess with above feature , BGP PIC provides only Hirerichal FIB so that it updates prefix very quickly compare to Flat table...

Am I correct in understanding ?  I am trying to understand that if I have eibpg multipath support in box , can I have convergance still faster if say no BGP PIC Edge in software and if not what BGP PIC Edge will provide me further.....

Sorry if I am askign same question again but i am really trying to understand BGP PIC Edge advantage as other vendor doesn't support this and claim that we still support multipath and you can achieve same convergance...like Juniper :-)

Regards,

Chintan