Re: Advancing the Protocol and Morin Drafts

[Robert Raszuk <[e-mail]>]

Hi Maria,

PIM Join latency would be in unicast world comparable to single prefix 
propagation by BGP.

Configuration I was talking about is not special .. but just correct :).

So in your network are you saying that single VPN prefix propagation in 
loaded production takes order of seconds from the moment one CE sends it 
to it's closest PE and the other CEs installs it in the RIB/FIB ? That's 
way too much I agree.

The only possible cause I could think of would be MRAI timer but most 
implementation I am familiar with have either this default to zero or it 
possible to be set as zero by single knob.

Perhaps what you are saying is that you may not have control over CEs 
and some BGP code there by default applies MRAI while PIM does not have 
such concept. That scenario indeed would be the cause of the delay for 
BGP propagation and not for PIM even if your core's BGP configuration 
would be already optimal.

Thx,
R.

> Hi Robert,
> we can achieve < 150 ms PIM Join latency (US cost-to-cost) without any
> special configuration. This yet has to be demonstrated with any other
> protocol in a loaded production environment...
> 
> Maria