Re: [Idr] Adoption Call on DT document (draft-ietf-idr-bgp-autoconf-considerations-00.txt) [3/10/2021 to 3/31/2021]

Hi Sue,

Ad 1-3)

As a contributor I do not support progressing the document in the
current form.

In my opinion set of requirements is too broad to what is required to
provide a simple discovery of BGP peer candidate. IMO single bit in LLDP
message or as proposed either BGP OPEN to link multicast or sort message to
link multicast should suffice.

Number of requirements already duplicate what is exchanged between peers in
BGP OPEN.

Is a BGP speaker capable of autodiscovery supposed now to filter peers with
the same BGP_ID ?

Why are we stating something against RFC5082 which says:

   "If, however, dynamic negotiation of GTSM support
   is necessary, protocol messages used for such negotiation
*MUST be   authenticated* using other security mechanisms to prevent DoS
attacks.

Why do we need to send a device role and what RFC defines the device role ?
You can't put requirements in place something which is today
nowhere deployed.

Support AFI/SAFI is already covered by BGP Capability negotiation. Do we
need to repeat those procedures here ?

Requirements list IPSec as one of transport security options .. and now
what if peer exchange such requirement ? Is this optional for a candidate
peer or mandatory ? What if list of transport security requirements is
exchanged ? Is there no requirement for preference order ?

Ad 4)

We have been working in IDR for many years on
https://tools.ietf.org/html/draft-raszuk-idr-bgp-auto-discovery-06

So now perhaps more valuable would be to review this work instead of
creating a requirements document ?

Many thx,
Robert

On Wed, Mar 10, 2021 at 7:03 PM Susan Hares <shares@ndzh.com> wrote:

> Greetings:
>
>
>
> This is an adoption/feedback call for the DT team document:
>
>
> https://datatracker.ietf.org/doc/draft-ietf-idr-bgp-autoconf-considerations/
>
>
>
> This call for comments will run for 3 weeks.
>
>
>
> In your feedback please let us know:
>
>
>
> 1) do you think this represents a good set of requirements for bgp
> autoconf protocol in the data center?
>
>
>
> 2) Do you think the document should remove any requirements or add an
> requirements?
>
>
>
> 3) Do you think the review of the protocols should be moved to another
> document?
>
>
>
> 4) Should a IDR DT create the requirements for non-Data Center deployments
> prior to starting work on a BGP auto-configuration protocol? If so, should
> a DT start these requirements in parallel?
>
>
>
> Cheers, Sue Hares
>
>
> _______________________________________________
> Idr mailing list
> Idr@ietf.org
> https://www.ietf.org/mailman/listinfo/idr
>