Re: [eppext] New Version Notification for draft-brown-epp-fees-06.txt

["Gould, James" <[e-mail]>]

Pat & Roger,

Let me clarify from two different perspectives:

Idealist:
The fee information really doesn’t follow the semantics of a check, but follows the semantics of an info.  The check command is setup as a container of boolean checks of the existence or availability of something (domain, claim, contact identifier, etc.).  The info command provides the query information for general information.  The fee extension is not asking for the existence or availability of something, but is asking for fee information given query inputs.  The query inputs may be invalid, so one of my questions is what should the server return for each of the invalid inputs (objID, objURI, objID, element, currency, command, and period)?  The check command should be simple and fast, where extending it with additional query or separate check information runs the risk of making it complex and slow.

Pragmatist:
If a fee is needed for every domain check, then we need a mechanism to support fee information as an extension to check.  To keep it a true extension, the only elements needed in the <fee:check> element should be <fee:currency>, <fee:command>, and <fee:period> that apply across all of the check identifiers (<domain:name>) included under the check command (<domain:check>).  This would result in a check response including a <fee:chkData> element that matches what is currently defined in draft-brown-epp-fees-06, except the <fee:objID> elements would always correspond to a response identifier (<domain:name>).  If there is invalid input provided with <fee:currency>, <fee:command>, or <fee:period>, the check command should fail with a specified EPP error code (my guess is 2306, since invalid input would be based on server policy).

I’m assuming that the pragmatist is favorable over the idealist, but the idealist in me has a real concern of overloading the check command / response as a pattern to combine commands / responses.


—


JG


[cid:[e-mail]]

James Gould
Distinguished Engineer
[e-mail]

703-948-3271
12061 Bluemont Way
Reston, VA 20190

VerisignInc.com<http://VerisignInc.com>

On Dec 4, 2015, at 11:29 AM, Pat Moroney <[e-mail]<mailto:[e-mail]>> wrote:

Hi all,

I have to agree with Roger. The vast majority of check commands that we send are triggered by a customer searching on our website. To create the search results we need both the availability and the fee information in order to calculate the pricing. We can currently get that with one check command, but if it is split up our command volume will almost double.

Thanks,
-Pat

On Fri, Dec 4, 2015 at 9:20 AM Roger D Carney <[e-mail]<mailto:[e-mail]>> wrote:
Good Morning,

Interesting idea Jim. Though not completely disagreeing with your premise that the check and fee have different purposes, let me play devil’s advocate here with some knowledge on the separate claims check in mind as well. Different individual purposes maybe but when looking at the varied data flow scenarios (how is this data used and presented to the end customer), I believe most clients would see the purposes aligned well.

I have heard from many people with the same question about claims: “Why do we have to make two calls to see if a domain has a claim associated to it, can’t they just tell us in the check?”

I am guessing the same question will come up if it is a separate call for fee information as well. People may make the argument that the additional claims check is only needed for a specific period of time (except for those doing indefinite claims). I think everyone would agree this same argument does not apply for fees and as we see more and more varied wholesale models I think an integrated call is more natural and becomes more necessary.

Again, as these concepts align well from a flow perspective and as the probability of expanding wholesaling models increases, I believe most clients would prefer (demand) one call over multiple calls.


Thanks
Roger


From: EppExt [mailto:[e-mail]<mailto:[e-mail]>] On Behalf Of Gould, James
Sent: Friday, December 04, 2015 7:37 AM
To: Gavin Brown
Cc: [e-mail]<mailto:[e-mail]>
Subject: Re: [eppext] New Version Notification for draft-brown-epp-fees-06.txt

Gavin,

In reviewing draft-brown-epp-fees-06, I have the following feedback:


  1.  The fee check extension is really not a good fit for extension of the available check.  They have different purposes and there is really no relationship between the two, where the availability check can include a completely different set of domain names from getting the fee extension.  My recommendation is to separate the two and define the fee check as a new verb similar to the claims check in draft-ietf-eppext-launchphase.  Another option is truly extend the object identifiers (e.g. domain:names) in the check command with a set of fee attributes to get the fee information for.  Have the extension only specify the command and the optional currency and fee that applies to all of the names in the availability check command.  The response would include the same list of object identifiers in the extension, but it would be a one-to-one relationship.  In this way there is no need for the “objURI” attribute, the <fee:objID> element, and <fee:objID> “element” attribute in the check command, since the object and identifying element is already defined by the availability check object that is being extended.
  2.  What should be returned when the client passes invalid data in the check command for the the “objURI” attribute (e.g. “urn:made:up:uri”), the <fee:objID> element (e.g. invalid domain name), the “element” attribute of <fee:objID> (e.g. “madeup”), the <fee:currency> element (e.g. “MUP”), and the <fee:period> element (e.g. “99" when max is “10”)?  An invalid <fee:command> is covered by not returning the fee information, but it’s unclear whether the server should return an error or not return the fee information for other invalid data.  To follow the semantics of a check command and response, my recommendation is to add an “avail” attribute to either <fee:cd> or <fee:objID> to explicitly indicate that the fee information is available along with an optional <fee:reason> to indicate the reason that the fee information is not available.  The reason the fee information is not available for a particular object could be due to the passing of invalid input in the check command <fee:object> element.  If you went with a single set of fee elements in the check command that is applied to all of the object identifiers in the availability check, then an invalid fee element could result in a failure of the check command itself instead of being an error (e.g. not available) on a per domain basis.
  3.  Since the check response supports returning a “0.00” value to indicate no fee, doesn’t make sense to do the same with all transform command responses?


—

JG

[cid:image001.png@01D12E75.B8643000]

James Gould
Distinguished Engineer
[e-mail]<http://[e-mail]/>

703-948-3271
12061 Bluemont Way
Reston, VA 20190

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

On Nov 4, 2015, at 5:58 PM, Gavin Brown <[e-mail]<mailto:[e-mail]>> wrote:

Changes from 05 to 06:

  1.  The specification is now object-agnostic, but works with RFC5731
      [RFC5731] domains by default.

  2.  Renamed the <fee:domain> element to <fee:object>.  Added the
      "objURI" attribute.

  3.  Removed the default value for the "refundable" attribute of
      <fee:fee> elements, and added text about how clients should
      handle such cases.  Added similar text to the documentation of
      the "grace-period" attribute.

  4.  Removed references to the defunct <info> command syntax.

  5.  "MUST" requirements regarding documentation have been changed to
      "must".

  6.  Created separate "Correlation between Refundability and Grace
      Periods" section describing how the "refundable" and "grace-
      period" attributes work together.

Feedback welcome as always!

-------- Forwarded Message --------
Subject: New Version Notification for draft-brown-epp-fees-06.txt
Date: Wed, 04 Nov 2015 14:52:30 -0800
From: [e-mail]<mailto:[e-mail]>
To: Gavin Brown <[e-mail]<mailto:[e-mail]>>, Jothan Frakes
<[e-mail]<mailto:[e-mail]>>


A new version of I-D, draft-brown-epp-fees-06.txt
has been successfully submitted by Gavin Brown and posted to the
IETF repository.

Name: draft-brown-epp-fees
Revision: 06
Title: Registry Fee Extension for the Extensible Provisioning Protocol
(EPP)
Document date: 2015-11-04
Group: Individual Submission
Pages: 36
URL:
https://www.ietf.org/internet-drafts/draft-brown-epp-fees-06.txt
Status:         https://datatracker.ietf.org/doc/draft-brown-epp-fees/
Htmlized:       https://tools.ietf.org/html/draft-brown-epp-fees-06
Diff:           https://www.ietf.org/rfcdiff?url2=draft-brown-epp-fees-06

Abstract:
  This document describes an Extensible Provisioning Protocol (EPP)
  extension mapping for registry fees.





Please note that it may take a couple of minutes from the time of submission
until the htmlized version and diff are available at tools.ietf.org<http://tools.ietf.org/>.

The IETF Secretariat


--
Gavin Brown
Chief Technology Officer
CentralNic Group plc (LSE:CNIC)
Innovative, Reliable and Flexible Registry Services
for ccTLD, gTLD and private domain name registries
https://www.centralnic.com/

CentralNic Group plc is a company registered in England and Wales with
company number 8576358. Registered Offices: 35-39 Moorgate, London,
EC2R 6AR.



_______________________________________________
EppExt mailing list
[e-mail]<mailto:[e-mail]>
https://www.ietf.org/mailman/listinfo/eppext

_______________________________________________
EppExt mailing list
[e-mail]<mailto:[e-mail]>
https://www.ietf.org/mailman/listinfo/eppext
--
-Pat Moroney
Sr. Software Engineer
Name.com<http://Name.com>
http://www.youtube.com/watch?v=V1GKGXXF12c
720-663-0025
<image001.png><image001.png>_______________________________________________
EppExt mailing list
[e-mail]<mailto:[e-mail]>
https://www.ietf.org/mailman/listinfo/eppext