Marco,


The issue is associated with a transition state when one side of the connection 
(registrar or registry) supports EAI while the other does not, the email 
element is required by the protocol, and the only value available is an EAI 
address.  The EAI-supporting side can’t pass the EAI value since is known that 
the other side doesn’t support it based on the EPP greeting or login services.  
The proposal in the draft is to pass a predefined placeholder value 
([email protected]<mailto:[email protected]>) to signal the existence 
of an EAI value to a non-EAI supporting party, which is temporary until both 
parties support EAI.  This is a corner case that needs to be considered until 
both sides support EAI.  Let's consider some specific examples:



  1.  EPP Contact <create> Command 
(https://datatracker.ietf.org/doc/html/rfc5733#section-3.2.1) – The 
<contact:email> element is required, which poses an issue when the registrar 
supports EAI and the registry does not support EAI.  What should the registrar 
pass to the non-EAI supporting registry?  Collecting an ASCII and EAI email 
address for this case may be logical, but it’s counter to the goal of 
supporting EAI, since the registries that the registrar interfaces with can 
have a mix of EAI support.
  2.  EPP Contact <info> Response 
(https://datatracker.ietf.org/doc/html/rfc5733#section-3.1.2) - The 
<contact:email> element is required, which poses an issue when the registry 
supports EAI and the querying registrar does not support EAI.  The querying 
registrar could be any registrar and not just the sponsoring registrar, where 
there is only a single email value in the registry that is an EAI address for 
this use case.  What should the registry return to the non-EAI supporting 
registrar?  This case is even more challenging then case #1 since registry has 
a single value and there can be a mix of EAI supporting querying registrars.  
One alternative is to return an error response (e.g., 2308 “data management 
policy violation”) instead of a successful response with a predefined 
placeholder value,  but that seems harsh to me.



I believe the use of a predefined placeholder value is the best approach to 
handle this transition corner case in the protocol, but I would like to hear 
viable alternates from the working group to incorporate into the draft.



Thanks,



--



JG







James Gould

Fellow Engineer

[email protected] 
<applewebdata://13890C55-AAE8-4BF3-A6CE-B4BA42740803/[email protected]>



703-948-3271

12061 Bluemont Way

Reston, VA 20190



Verisign.com <http://verisigninc.com/>



On 8/2/21, 9:12 AM, "regext on behalf of InterNetX - Marco Schrieck" 
<[email protected] on behalf of [email protected]> wrote:



    Hi All,



    I am completely with Ulrich.



    >From policy side and also from aspect of clean data. Using dummy values

    in objects was always a bad idea.



    And even in GTLD space, the registrar usually have requirements to

    verify the contacts. So not only registry must support the EAI also the

    registrar have to do it.



    Regards

    Marco





    Am 02.08.21 um 14:47 schrieb Dmitry Belyavsky:

    > Hello Regext,

    >

    > First, many thanks to James for his presentation of the draft update.

    > As it was suggested, let's move the discussion to the ML.

    >

    > As usual with EAI, we have a chicken-and-egg problem. I agree with

    > Ulrich, that using a placeholder instead of a regular address doesn't

    > leave a Registry any option to communicate with a domain administrator

    > via email. But enforcing the domain administrator using a valid ASCII

    > address effectively stops it from moving to EAI. So placeholder is

    > implied to be a temporary measure until EAI is not widely supported. I

    > understand that the support of EAI will be different in different CC

    > TLDs, but I think it's worth supporting such addresses in the GTLDs.

    >

    > I agree that the question is about the intersection of the technical

    > aspects (where it's reasonable to provide the placeholder to make the

    > object created), legal stuff (where the valid address may be required)

    > and policy stuff (that imply the EAI addresses should be treated as just

    > email addresses).

    >

    > --

    > SY, Dmitry Belyavsky

    >

    > _______________________________________________

    > regext mailing list

    > [email protected]

    > 
https://secure-web.cisco.com/1knrAJeJSAijatcnZRFL2-q2sa9iAqop6c6aXEBkL5u98WhGqbukYOUR__gp59Q-GPk2aFY1BoKzb6RlNpmY12zqTx4AeARpNRf5ZNCoFRN9Nedd7nMTqIDLZrfj5EITKSKGL_pUoITxWiuxLvHGVhT7Hp_XNuKkpIe9HRifjofrBMrGzk3MoG1Fd6Of8Iw4VRUlKNTt5fmkSfKcfkwApqjGCFzFx86_1QIEGwora_SXwhwOpVIZb1Ter2ZZdWvy0N-TnIZceYeRBalZr7K1nnA/https%3A%2F%2Fwww.ietf.org%2Fmailman%2Flistinfo%2Fregext

    >



    _______________________________________________

    regext mailing list

    [email protected]

    
https://secure-web.cisco.com/1knrAJeJSAijatcnZRFL2-q2sa9iAqop6c6aXEBkL5u98WhGqbukYOUR__gp59Q-GPk2aFY1BoKzb6RlNpmY12zqTx4AeARpNRf5ZNCoFRN9Nedd7nMTqIDLZrfj5EITKSKGL_pUoITxWiuxLvHGVhT7Hp_XNuKkpIe9HRifjofrBMrGzk3MoG1Fd6Of8Iw4VRUlKNTt5fmkSfKcfkwApqjGCFzFx86_1QIEGwora_SXwhwOpVIZb1Ter2ZZdWvy0N-TnIZceYeRBalZr7K1nnA/https%3A%2F%2Fwww.ietf.org%2Fmailman%2Flistinfo%2Fregext
_______________________________________________
regext mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/regext

Reply via email to