Thanks for the comments, Sam.  More below.

-Scott-

> -----Original Message-----
> From: Samuel Weiler [mailto:[EMAIL PROTECTED] 
> Sent: Thursday, August 26, 2004 11:34 AM
> To: [EMAIL PROTECTED]
> Cc: Scott Hollenbeck
> Subject: Re: [dnsop] DNSSEC and EPP Review Request
> 
> 
> Comments on draft-hollenbeck-epp-secdns-03.txt:
> 
> Perhaps I'm missing something, but Section 2.2 and Section 7 (security
> considerations) both refer to sending signatures, presumably over EPP,
> yet I see no mention of those data elements in the schema.  Which is
> correct?  I suspect it's sufficient to say "you need to authenticate
> this somehow", which section 7 does.

Section 4, this element:

<element name="sigVal" type="secDNS:keySigType"/>

> In section 3.1.2, the nxt attribute implies the use of opt-in.  This
> should probably be removed or discussed only as a "future extension"
> bit.  Also s/NXT/NSEC/.

Thanks for catching this.  It's an artifact of older thinking.

> There are elements for start/end dates on publication, but nothing for
> sig lifetime.  It may be useful for a client to be able to request a
> sig lifetime much shorter than the EXPECTED key publication interval.
> If the start/end dates are meant to be that, this needs to be
> clarified.

OK, what's really useful from a deployment perspective?  Are the start-end
dates good enough, or is something else needed?  What are the implications
of the sig lifetime and start-end dates being different?  I'm open to ideas.

> The startDate/endDate elements appear to be part of dsData but not key
> (they're in keyData, but not key).  The rem element requires either a
> key or dsData.  Hence, rem'ing a dsData appears to require sending
> those dates.  It may not be wise to require the registrar to remember
> what start/end dates it (or another registrar) originally requested in
> order to remove a DS record.

Clients don't have to remember the dates.  They can be retrieved via an
<info> query.  The text makes that clear, but I suppose the example could be
better.

> Also: can you remove something created
> as keyData by removing dsData? Or the other way around?  I'm surprised
> that rem requires recitation of the entire old dsData when chg does
> not.

Nope, there's a one-for-one relationship.

> Formats of data: the date field formats aren't specified.  For other
> fields, refer specificallly to the "presentation formats" in 2535 (and
> successor docs).  Consider whether you need any further escaping or
> other encoding for each field.

Date field formats are specified in the schema using the XML Schema
"dateTime" data type.  Other field formats *are* referenced to
draft-ietf-dnsext-dnssec-records.  Did I miss something somewhere?

> Update the doc to reference DNSSECbis.

Where (section 1 maybe?) and in what context?  There is no outdated
reference to replace.

-Scott-


.
dnsop resources:_____________________________________________________
web user interface: http://darkwing.uoregon.edu/~llynch/dnsop.html
mhonarc archive: http://darkwing.uoregon.edu/~llynch/dnsop/index.html

Reply via email to