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

[Lukas Tribus]

Hi,

On Wednesday, 10 December 2025, Gert Doering <gert at greenie.muc.de> wrote:

> Hi,
>
> On Wed, Dec 10, 2025 at 07:12:25PM +0100, Lukas Tribus via cisco-nsp wrote:
> > I don't understand.
> >
> > How is this supposed to work without transient loops and without
> > tunneling (label unicast/vpnv4), when your core has yet to converge?
> >
> > For this to work the internal routers you cross all would have to
> > converge their FIB at the same time at the same speed and in the same
> > order?
> >
> > What am I missing here?
>
> Without labels ("any sort of tunnels") you would see transient loops,
> but convergence would still be faster if the edge already has a candidate
> backup path - as compared to "send withdraw, wait for the other side(s)
> to process the withdraw, and send the new-best path back".  So it will
> be converged after "process withdraw, select new path".
>

Agreed and based on your other mail I now got your point about XR. Thanks
for clarifying.

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.

I also like to make a routing decison ideally only once, not even on the
egress box if possible (per next hop label allocation).

But those are "personal tastes".


Lukas