[RZERC] Possible scenarios to take up or turn down

[Howard Eland]

> On Aug 21, 2017, at 4:39 PM, Kim Davies <kim.davies at iana.org> wrote:
> 
> Hi folks,
> 
> Quoting Jim Reid and others on Monday August 21, 2017:
>>> 
>>> - Introduction of RRTYPEs not previously included in the root zone (this would cover DNSSEC in the past).
>>> - Changes which would cause a “significant” change in size to the root zone (Significant TBD).
>>> - Significant changes to the operation of the root zone, such as
>>> 	- Change in root zone operator
> 
> What do we envisage RZERC's role is in this scenario?

Perhaps a review of any new potential operator’s history as it pertains to DNS operations?  Impact study of how the new operator’s network coverage could potentially change query paths?

>  
> 
>>> 	- Change in DNSSEC key algorithm or size
>>> 	- KSK roll?
> 
> Assuming this KSK roll concludes in a normal fashion, it seems to me
> performing a key rollover will become perfunctory (i.e. rolling every x
> years). I don't see a KSK roll being elevated to this group unless
> it meets the other criteria above (i.e. new algorithm, possibly new
> size.)

What about emergency KSK rolls?  I’m not saying we bless a roll before it occurs, but should we look at the procedures for this type of activity to ensure they are in line with keeping the stability of the root zone?

> 
>>> - Changes in operational policy
>>> 	- Significant change in root DPS (Significant TBD)
>>> 
>>> 
>>> Possible list of topics the RZERC committee should decline to review:
>>> 
>>> - Daily “standard” changes to the root zone (Standard TBD).
>>> - Introduction of new TLDs into the root zone (subject to the size bullet above).
> 
> A couple of other scenarios I had in mind that don't reflect architectural
> changes to the root zone content itself:
> 
> - Support for additional algorithms and digest types for delegations.

I believe I mentioned this above for the root zone, but yes, the same could be said for delegations.

> - Upgrades to the software used to manage the root zone and root zone workflow
>  (except as it pertains to other characteristics deemed to be in-scope)

Not sure I want to do work every time there’s a patch to an operator’s software.  Would that also include network gear upgrades?  This seems intractable to me.

> - Revisions to the technical check criteria for the root zone.

+1 

> - Revisions to the authentication mechanisms for TLD managers.

Not sure exactly what you mean on this last one.

> 
>> One thing that I think could/should be added to the list is getting
>> a root key repository outside the USA. Suppose US airspace gets closed
>> down. Or immigration policies make it impractical for foreigners to
>> serve as TCRs.
> 
> What would be the scope of RZERC's consideration of such a change? Where
> the root management partners keep their physical facilities seems to me
> beyond "architectural changes to the content of the DNS root zone”.

I’m concerned that this moves us more into Layer 9 - are we going to look into International policy changes?

> 
>> I assume that RZERC is not going to do any work on whatever ends up
>> on the list we compile. Until of course someone approaches us with a
>> proposal that’s in scope for RZERC.
> 
> I think that's correct. What we discussed on the call was having illustrative
> scenarios that would inform thinking about what is in-scope and out-of-scope.

+1

-Howard

> 
> kim