Re: [ietf-provreg] RFC updates - roid length restriction

[James Gould <jgould@verisign.com>]

Alex,

Are you referring to the length of the contact client id or the length of
the roid?  The clIDType is defined as:

  <simpleType name="clIDType">

        <restriction base="token">

            <minLength value="3"/>

            <maxLength value="16"/>

        </restriction>

    </simpleType>



The roidType is defined as:

  <simpleType name="roidType">

        <restriction base="token">

        <pattern value="(\w|_){1,80}-\w{1,8}"/>

        </restriction>

   </simpleType>



The EPP roidType supports your format, but the client id's would have to be
between 3 and 16 characters.  We didn't have a migration issue to deal with.
My recommendation is to use the prefix values (AMFR12011011) or something
else as the client id's and use the full object id (AMFR12011011-NICAT) as
the roid.  


-- 

JG 

James F. Gould
VeriSign Naming and Directory Services
jgould@verisign.com

This message is intended for the use of the individual or entity to which it
is addressed, and may contain information that is privileged, confidential
and exempt from disclosure under applicable law. Any unauthorized use,
distribution, or disclosure is strictly prohibited. If you have received
this message in error, please notify sender immediately and destroy/delete
the original transmission


> From: Alexander Mayrhofer <axelm@nic.at>
> Date: Thu, 29 Sep 2005 16:27:40 +0200
> To: <ietf-provreg@cafax.se>
> Subject: [ietf-provreg] RFC updates - roid length restriction
> 
> 
> We are currently evaluating EPP to replace our email based provisioning
> system in our ccTLD registry (.at)
> 
> One of the (surprisingly few remaining) issues we encountered is that
> handles for our contact objects extend beyond the 16-character limit of EPP
> repository object id's (we have handles like "AMFR12011011-NICAT". We don't
> want to change or even drop the suffix on our object identifiers (because
> all of our registrars would have to do the same in their internal databases)
> so our only option now seems to go for either an extension of some kind, or
> to completely replace contact & domain mapping with our local mappings.
> 
> The different roid length would then be the only essential difference
> between stock and local mappings, so we'd like to avoid that.
> 
> Questions:
> 
> - Did any other registry operators already encounter this, or are we the
> only ones who use "oversized" handles (i can't believe that)?
> - Should therefore the length restriction be extended (to eg. 32 characters)
> in the "refurbished" drafts to safely avoid such problems?
> 
> any comments appreciated.
> 
> Alex Mayrhofer
> nic.at
>