Re: Last Call: <draft-iab-2870bis-01.txt> (DNS Root Name Service Protocol and Deployment Requirements) to Best Current Practice

["Carlos M. Martinez" <[e-mail]>]

David,


On 30/05/2014 14:17, David Conrad wrote:
> Carlos,
> 
> On May 30, 2014, at 8:56 AM, Carlos M. Martinez <[e-mail]> wrote:
>> If you
>> agree to provide a service that the whole internet depends on, then you
>> need to comply with a few requirements.
> 
> That might be nice, but that isn't how the root server system works.

I understand. However, my point is that this shouldn't be a
consideration when setting requirements. The requirements need to match
what the Internet needs now and will need for the next 5, 10, 15 years.

The discussion about how to fulfill the requirements starts when the
requirements are set.

> 
> The root server operators provide a voluntary service for their own reasons using their own resources. History has shown that no one other than the folks that pay to root server operators' bills gets to impose requirements on how that service is provided. The IETF, IESG, IAB, ICANN, and/or random people off the street can make suggestions on how that service can be provided, but there should be no illusions about whether those suggestions are going to be followed.

I do thank them for doing that. Some of them have been doing this for
decades. I'm pretty sure that some of them actually took a risk
accepting this responsibility back at the time. They deserve praise.

However, my point above remains. Requirement setting should be based on
actual current and future needs, not on the hows and whys things were
managed in the past.

> 
>> If you can't / won't, well... you can opt out.
> 
> Ignoring for the moment the fact that there still is no succession plan if a root server actually were to opt out, there simply is no incentive for a root server operator to opt out. As such, suggesting that as an option is ... of debatable value.

The fact that there is no succession plan is actually something that is
broken in the IG arena. IMO, it's not the IETF's fault, nor the IETF's
responsibility to correct; but it's broken nonetheless.

This fact was noticed during NetMundial in at least two meetings I
attended. It was awkward. I think 'We the Internet as we know it' need
to pay attention to this and fix it before 'others' think they need to
fix it for us.

> 
>> But past failures in enforcement
>> should not deter the IETF of setting the requirements the IETF deems
>> necessary for the correct operation of the Internet.
> 
> I guess I just don't see the point. I suppose it doesn't really do any harm, but as Patrik states, it isn't going to have any impact so why bother? 

Maybe I like having guidance. A requirements document would provide
useful guidance. Maybe it should not be a BCP, I don't know. But having
written requirements provides useful basis for further discussion on how
to implement them.

My main fear is actually what you perfectly described in another email.
If we leave requirements like IPv6 and anycast and feature X, out, and
leave them just to the good will of some organizations, well, we risk
the scenario where some might say 'why should I do it if no one is doing
it'.

regards,

~Carlos

> 
> Regards,
> -drc
>