[c-nsp] bgp PIC Edge & neighbor shutdown

[James Bensley]

On 15 August 2017 at 10:00,  <adamv0025 at netconsultings.com> wrote:
>> Vladimir Troitskiy
>> Sent: Friday, August 11, 2017 10:59 AM
>>
>> Hi all,
>>
>> BGP PIC Edge on an egress PE tracks PE-CE interface and switches to backup
>> path immediately when link to the CE goes down. But what if I manually
>> shutdown the PE-CE eBGP session? Does it trigger immediate CEF update on
>> the egress PE (ASR9k)?
>> --
> Let's do a little thought experiment to see what might be going on under the
> hood.

Yes a thought experiment is basically what I was getting at in my post.

> Let's assume you shut down the session on local (A)end.
> If BGP session is shut down, then BGP on local (A) end will immediately
> withdraw all prefixes learned via that session, (BGP on remote (B) end will
> terminate the session after receiving BGP Cease NOTIFICATION).
>  -now I think what you're really asking is whether this is done in a PIC way
> (i.e. triggers artificial BGP-NH invalidation thus all prefixes at once) or
> not (BGP withdraws prefix by prefix and notifies RIB for each), and my
> answer is I don't know.

I was approaching it from the interface level shutdown, not the BGP
level shutdown.

However, with RIB updates in BGP (whether prefixes are withdrawn one
by one or all together) the next step I believe is to share these
route changes with the central device RIB. I would have thought
(hoped?) that irrelevant of weather BGP RIB to “main” RIB API says
"Hey, ALL these prefixes are no longer reachable via A.B.C.D NH
(source eBGP)" or it says "Hey, THIS prefix only via A.B.C.D NH is
gone (source eBGP)" or it says "Hey, A.B.C.D NH address (and thus any
routes via this NH) is gone (source eBGP)" - in any case, the RIB
would then have to look at any other routes offered to it from BGP or
other routing protocols, if a backup is found, then install that into
FIB and program the updates into hardware CEF tables.

Going from the interface level upwards, with BGP PIC Core and Edge
configured, we have (almost) instantaneous forwarding plane recovery,
having pre-computed a backup path and having it already programmed
into CEF and marked as a backup. When there is a physical link down
and RIB notifies the FIB the backup path is pushed as the main path
etc.

When the next hop is invalidated by a protocol update (coming from the
top down) BGP RIB says to central FIB “this NH is gone” and central
RIB has an alternate paths via another eBGP peer (which would be the
PIC path used during an interface/link down failure) does the central
RIB say to FIB and on to CEF hardware “switch to known backup path” or
does it re-compute the paths because this was a control plane
triggered update not a forwarding plane triggered update?

Cheers,
James.