[c-nsp] Best Practices for quickly removing routes when BGP peer drops

[Gert Doering]

Hi,

On Wed, Dec 10, 2025 at 11:26:20PM +0100, Lukas Tribus wrote:
> I'm interested to hear why folks don't likel labeling internet traffic.
> 
> I always l liked it, indeed I have a real distaste for programming the full
> table to the FIB on core only boxes.
> 
> Perhaps thats because I used real slow FIB convergence boxes for a long
> time.

For us, it's a bit of "historic baggage".  The network is really old,
and while all components and links have been upgraded numerous times, 
some fundamental decisions have never been changed.

MPLS was added as "an extra feature" at some point in time, to enable
L2 circuits and L3 VPNs.  At that point we wanted to keep technologies
distinct, mostly due to "lack of experience with MPLS", so Internet (v4+v6)
was kept as native/unlabeled.  Which had the benefit that people trying
to debug "Internet" did not need to understand or care about MPLS - and 
the drawbacks of "the core needs to converge" and everything you say.

Nowadays, MPLS L3 VPNs have gone again (and are replaced by SD-WAN style
meshes on managed firewall products), so operational experience with MPLS
is still low.  EVPN/VXLAN has been added to the portfolio for L2 ELAN
products - something extra people need to learn, less brain capacity for
MPLS again...

So for us it still is "the network is too small, there's not enough 
engineers to fully understand and operate *and troubleshoot* all the
technologies", coupled with some technology limitations (Arista 7050SX3
series not admitting they could do some MPLS P).  Meh.


... this all said, we've been quite happy with the operational stability
of our peering and transit links, so the whole topic of "fast covergence
if something breaks" has been very low on the priority list for the last
few years.  For maintenance, we use GSHUT these days, which really nicely
converges everything to new paths before the actual link shutdown...

gert
-- 
"If was one thing all people took for granted, was conviction that if you 
 feed honest figures into a computer, honest figures come out. Never doubted 
 it myself till I met a computer with a sense of humor."
                             Robert A. Heinlein, The Moon is a Harsh Mistress

Gert Doering - Munich, Germany                             gert at greenie.muc.de
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 659 bytes
Desc: not available
URL: <https://puck.nether.net/pipermail/cisco-nsp/attachments/20251211/3b1382ea/attachment.sig>