[c-nsp] BGP ORR - experiences

[Phil Bedard]

Hi Darko,

If you mean the 9001, that router was end of life quite some time ago and of course the software running on it stopped being updated a couple years ago.   We (Cisco, I work for them)  have a couple very large carriers using XR based RRs with ORR, but it’s on something newer than a 9001.

Phil

From: Darko P <011.darko at gmail.com>
Date: Thursday, October 23, 2025 at 2:07 PM
To: Phil Bedard <philxor at gmail.com>
Cc: Mark Tinka <mark at tinka.africa>, Saku Ytti <saku at ytti.fi>, cisco-nsp at puck.nether.net <cisco-nsp at puck.nether.net>
Subject: Re: [c-nsp] BGP ORR - experiences

Thank you for the feedback!

Cisco responded that ORR is not supported on the 9001 due to numerous bugs. It's a software thingy.
They noted some 'defects' that have been fixed beginning with version 7.5.x.

Anyway, after Cisco's lengthy explanation, I am now leaning more towards Juniper or Nokia for the RRs.

We're kind of in a situation where we have over 200 PoPs, and ORR would solve many of our problems in the mid-term. It's easier to buy a couple of new RRs than several hundred routers in this economy. :)

Cheers,

Darko

On Tue, Oct 21, 2025 at 12:50 AM Phil Bedard <philxor at gmail.com<mailto:philxor at gmail.com>> wrote:
A bit late but we (Cisco) do have some folks running ORR but not many.   Recently I worked with a large provider and in the end they decided to just distribute RRs across their network instead.  I’m not aware of any limitations with the ASR9K specifically in regards to ORR, there is no silicon dependence with ORR in IOS-XR.   It’s what we would call a platform-independent function.

However, we typically recommend folks look at the RR appliance or XRv9K/xRD for RR use cases since the general CPUs are much more powerful than the ones we put into a router RP.

Thanks,
Phil

From: cisco-nsp <cisco-nsp-bounces at puck.nether.net<mailto:cisco-nsp-bounces at puck.nether.net>> on behalf of Mark Tinka via cisco-nsp <cisco-nsp at puck.nether.net<mailto:cisco-nsp at puck.nether.net>>
Date: Thursday, October 16, 2025 at 12:54 AM
To: Saku Ytti <saku at ytti.fi<mailto:saku at ytti.fi>>, Darko P <011.darko at gmail.com<mailto:011.darko at gmail.com>>
Cc: cisco-nsp at puck.nether.net<mailto:cisco-nsp at puck.nether.net> <cisco-nsp at puck.nether.net<mailto:cisco-nsp at puck.nether.net>>
Subject: Re: [c-nsp] BGP ORR - experiences



On 15/10/2025 09:23, Saku Ytti via cisco-nsp wrote:

> We use ORR and add-path 3 (to enable ECMP and to have backup) with
> about 40 ORR perspectives and +15M RIB. But this is Junos running on
> compute.
>
> You really want compute, not actual router hardware as the CPU gains
> from modern compute are very significant.

In my IP days (up until about 2 years ago), we ran Add-Paths on a Cisco
CSR1000v (was in the process of transitioning to the 8000v) route
reflector with Junos clients.

Worked very well that we did not need ORR, having up to 6 paths per
client with BGP Multipath enabled as well.

Ate a lot of CPU and RAM on the clients, but I've since heard that those
RE's were upgraded, so no more concern.

Being able to load balance traffic to the same AS across different
cities/regions was very sweet. Opened up some new commercial
opportunities the competition could not support.

Mark.
_______________________________________________
cisco-nsp mailing list  cisco-nsp at puck.nether.net<mailto:cisco-nsp at puck.nether.net>
https://puck.nether.net/mailman/listinfo/cisco-nsp
archive at http://puck.nether.net/pipermail/cisco-nsp/