Re: [Idr] draft-white-bgpbgp-00.txt

Hi Russ,

I think exploring new ideas is always great providing they bring real
value at right price.

So far you keep complaining that things which we carry in BGP have
nothing to do with destination based routing.

Then perhaps a better option is to issue a short one page draft
renaming Network Layer Reachability Information "NLRI" field to
"Database Key"  field - the advantages are that your concern will get
addressed yet multiple implementations, deployments and operational
practices will remain unchanged.

- - -

Meanwhile if we are at concerns regarding BGP I think we have two
serious problems in BGP today:

* dynamic capabilites
* multisessions

Both are either not supported even partially in most implementations
(dynamic capabilites) or supported in a way that they could be
improved (recall my discussion with John in Stockholm IETF regarding
use of TI vs multisession for improving BGP session robustness).

Thx
R.

On Tue, Jul 8, 2014 at 9:07 PM, Russ White <russw@riw.us> wrote:
>
>> Whether this is "madness" or not is a matter of *opinion* (and each of us
> is
>> entitled to have one).
>
> Of course... I didn't mean to offend, just to say that we keep piling things
> into the AFI/SAFI that have no specific relation to destination based
> routing. "Stop the madness," is a saying, not a term of condemnation. :-)
>
>> existence proof (means deployment and operators) that the BGP AFI/SAFI
>> construction is Good Enough
>> (tm) for what one may consider as "nothing to do with routing".
>
> Sure -- but I also see the protocol becoming more complex over time because
> of this additional information. If there's a simpler way to carry this
> information, one that allows us to either stream the information out of band
> or in band, etc., then we should consider it, IMHO.
>
> :-)
>
> Russ
>
> _______________________________________________
> Idr mailing list
> Idr@ietf.org
> https://www.ietf.org/mailman/listinfo/idr