RE: BEEP Transport

["Hollenbeck, Scott" <shollenbeck@verisign.com>]

>-----Original Message-----
>From: Dave Crocker [mailto:dcrocker@brandenburg.com]
>Sent: Tuesday, September 04, 2001 1:12 PM
>To: Hollenbeck, Scott
>Cc: 'ietf-provreg@cafax.se'
>Subject: Re: BEEP Transport
>
>
>Scott,
>
>At 09:37 AM 9/4/2001, Hollenbeck, Scott wrote:
>>BEEP does not have an explicit dependency on TCP, though this draft
>>describes TCP connectivity.
>
>It certainly relies on a reliable, connection-oriented transport that 
>delivers data in the order sent.

Of course -- my only concern is that the current draft refers to 3081
without saying anything to explain the 3081 is one possible transport
mapping of many possibilities.  Someone might well ask "why not SCTP", so I
think it would be a good idea to be just a little more clear when describing
why TCP is required.

>
>>I believe it completely inappropriate for a WG draft to define a profile
>>using vendor-specific URIs.  I'm not aware of any IETF document that
>>describes IETF-standard BEEP profile definition procedures;
>
>What do you mean vendor-specific URI, and how is it relevant to discussion 
>of our use of BEEP?

http://www.neulevel.com/profiles/EPP/1.0

This is the URI used to define the profile in the draft.  Though it is noted
that this is temporary, it is clearly in the namespace of a particular
corporate entity.  If we are attempting to define a standard-track profile,
with software implementations that MUST reference a particular URI to
identify the profile, I think it would be better to use a more neutral
identifier.  Other BEEP profiles defined in IETF documents use identifiers
from the xml.resource.org namespace, for example.

<Scott/>