Re: [ietf-provreg] a new core command...?

[Klaus Malorny <Klaus.Malorny@knipp.de>]

Eugenio Pinto wrote:
> Dear James, Michael, Scott
> 
> Thanks for the comments.
> 
> I understand EPP wasn't designed to be a search protocol because the 
> idea was rejected since there was no space in a very big Registry like 
> COM/NET to such a funtionality.
> 
> We all know that EPP isn't anymore just a communication protocol used in 
> global and big Registries.
> ccTLD's are beeing interested on it and some premisses that let EPP core 
> developers to some decisions may now not be all so true...
> 
> 
> Just for the records:
> 
> EPP <renew> Command
> 
>   Renewal semantics do not apply to host objects, so there is no
>   mapping defined for the EPP <renew> command.
> 
> EPP <renew> Command
> 
>   Renewal semantics do not apply to contact objects, so there is no
>   mapping defined for the EPP <renew> command.
> 
> 
> So, nowadays we have a <renew> Transform Command used just for a type of 
> objects that is the "domain name mapping" because of a specific object 
> implementation that is the "domain name mapping"
> 
> I'm wondering why wasn't it defined an extension in the "domain name 
> mapping" instead of the introduction of the <renew> command in the core 
> protocol.
> 
> Let's see:
> 
> Suppose the development decision was the extension for the domain renew 
> command, we would be using a bigger message for the same functionality! 
> We would have a lost in performance.
> That's the kind of thing I want to prevent with the addition of the new 
> command.
> 
> Let's compare the <renew> command with a possible <search> command.
> 
> 1 - The renew command is actually used only for the domain name mapping
> 2 - A <search> command can be applied to every future object mapping
> 
> I know. We can't apply it when we are in a very large Registry 
> environment. But, who knows? Maybe if... Are you seeing what I mean?
> 
> There are a few commands that DNS.PT will not implement, like the 
> <delete> command (policy aspects).
> 
> And it's not because we don't need it that the command was not defined.
> 
> I think that in a global provisioning protocol, used in a different 
> variety of environments  a <search> command is much more useful then a 
> <renew> command. But it's just my humble opinion...
> 
> Is anyone "out there" that would like to use (or is actually using, by 
> means of extensions) an EPP command for searching objects stored in a 
> shared central repository?
> 
> That's the question...
> 
> (Please, reply to this email if you do)
> 
> Eugenio Pinto
> FCCN - DNS.PT
> 
> 



Hi Eugenio,

the fact is that while EPP has some good ideas and tries to keep things 
symmetrical and simple, it also has some nasty flaws, like the fixed command 
set, the fixed set of status values in the contact, host and domain mappings, 
the whole error reporting mechanism (fixed set of error codes, the error value 
mechanism which is flawed and mostly not understood and correctly supported by 
the registries) and so on.

In my humble opinion, EPP is too good that there would be an urgent need for a 
replacement, but at the same time also too bad to base any further protocol 
development upon it. Extensions as described in the RFCs are the way to go, but 
keep them simple and use them conservatively in order to do the registrars a 
favour. Don't make the mistakes like those that have been made with the 
so-called EPP implementation for the .eu domains, which is more or less an 
implementor's nightmare because of the large deviations from the established 
procedures.

Regards,

Klaus


___________________________________________________________________________
      |       |
      | knipp |                   Knipp  Medien und Kommunikation GmbH
       -------                           Technologiepark
                                         Martin-Schmeißer-Weg 9
      Dipl. Inf. Klaus Malorny           44227 Dortmund
      Klaus.Malorny@knipp.de             Tel. +49 231 9703 0