As I made the initial proposal for registries to implement validation based on the EPP standards in RFC5733 I’m sure it will be no surprise that I favor option 1.  I think RFC5733 is a well understood and defined standard that represents a nice middle ground as far as validation goes.

I recognize that there is an argument to be made for both less validation as suggested in option 2 or more validation as was expressed by some IRT members when the initial proposal was first discussed (for example there was concern that the address 1 field was not required).

If we do take the option 2 approach I request that we be clear on what is meant by “registries do not impose any checks on registration data…”  That’s fairly nebulous.  I assume we would still have all the same fields but for example under phone extension, currently only numeric values are accepted.  Would we remove this restriction and allow ASCII characters as well…. UTF8?  I would think we would still need to make the Contact ID and Auth Info fields required.  I would also want to keep the Max field lengths in place for the various fields as that has sizing implications on our systems.

I think it’s possible to sync the transition of new and existing registrations with either option but I think either way it will take longer for existing registrations.  I’m not opposed to syncing them up but I don’t think it makes sense to hold up one effort for the other.

I agree with Roger’s statement about having a discussion on what a bulk transfer is.

Thank you,

Marc

From: gnso-impl-thickwhois-rt-bounces@icann.org [mailto:gnso-impl-thickwhois-rt-bounces@icann.org] On Behalf Of Fabien Betremieux
Sent: Monday, April 04, 2016 6:48 PM
To: gnso-impl-thickwhois-rt@icann.org
Subject: [Gnso-impl-thickwhois-rt] Proposal for the Transition of Existing Registrations from Thin to Thick

Dear IRT members,

In our recent conference call, the IRT discussed the transition of existing registration from thin to thick. It is our understanding that two alternative approaches are emerging:

Option 1 - The registries impose some checks on the registration data before it can be accepted

• The initial proposal from the registries for such checks is based on EPP Standards (RFC 5733), with subsequent discussion of potential changes to such checks.
• The main drawback of this approach is that the transition would likely to last a considerable amount of time due to:

• The need for registrars to process a very significant amount of data (collectively) to ensure it would pass the registries’ checks
• The need for Staff and the IRT to gather findings from data analysis by registrars before they can define a realistic implementation timeline, which in itself would delay the definition of the implementation plan

Option 2 - The registries do not impose any checks on the registration data during the transition

• This is a proposal emerging from recent discussions, considering that the Policy Recommendation does not include data accuracy requirements and therefore is out of scope for this implementation
• The benefit of this approach is that it Is in scope with the policy recommendations, it reduces the implementation to a sizable bulk transfer of data, and it creates an opportunity to possibly synchronize the transition of new and existing registrations by defining a single cut-off date after which all registrations are thick.

Considering the outcome of the IRT’s meeting with the RrSG in Marrakech, and considering recent community comments on the time it is taking to implement the transition from thin to thick, we would like to propose that the IRT move forward with Option 2 as we believe it is the most applicable path forward.

We would like to gather IRT members thoughts on our proposal to move forward with Option 2. Your input would be appreciated by Friday 8 April COB at the latest, for discussion during our next IRT meeting, which we are planning to organize the following week.

Thank you in advance for your consideration

Best Regards