[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: IETF Last Call comments on draft-ietf-netconf-soap-06.txt
Please find my comments in line:
On 25-Nov-05, at 6:42 AM, Wijnen, Bert (Bert) wrote:
NetConf WG, and specifically author(s) of the SOAP document,
please review and comment.
Bert
-----Original Message-----
From: iesg-bounces@ietf.org [mailto:iesg-bounces@ietf.org]On Behalf
Of Mark Baker
Sent: Wednesday, November 23, 2005 21:37
To: iesg@ietf.org
Subject: NETCONF over SOAP and BCP 56
I'd like to take issue with a claim made in the NETCONF over SOAP
draft;
http://www.ietf.org/internet-drafts/draft-ietf-netconf-soap-06.txt
In section 2.4 regarding BCP 56, the draft says;
"Fundamentally, these concerns lie directly with SOAP over HTTP,
rather than the application of SOAP over HTTP to NETCONF."
That is incorrect. The advice of BCP 56 is relevant to any use of
HTTP.
I agree that BCP 56 is relevant to any use of HTTP; however
BCP 56 makes recommendations that are at odds with the widespread
practice of using HTTP as a universal tunneling protocol (however
dangerous it may be, it is widespread, and it is even the
expectation of many users of SOAP). Would the following wording be
preferred?
Fundamentally, many of these concerns lie directly with
common usage of SOAP over HTTP, rather than the application
of SOAP over HTTP to NETCONF.
For example, BCP 56 provides guidance about the use of port
numbers, security, and URI schemes, none of which SOAP (1.1 or 1.2)
take any position on. Moreover, the SOAP specifications don't
require that HTTP be used as a tunnel; they fully support the use
of SOAP as an extension to the HTTP processing model.
IMHO, either the draft needs to defend its disregard of many of the
recommendations of BCP 56, or it needs to accomodate them as best
it can. I expect this will be particularly difficult given that
NETCONF is itself an application protocol, but at the very least I
think the draft should recommend or require that port 80 not be
used, or that a new HTTP method be used (viz a viz BCP 56 section
6), so as to separate NETCONF traffic from Web traffic.
Requiring a new HTTP method would make NETCONF incompatible with
existing SOAP implementations, but there is no difficulty with
requiring a specific port. The draft currently contains the
following text:
It is also possible to respond to the concern on the
re-use of port 80. A NETCONF SOAP service SHOULD be offered
over a new standard port for NETCONF over SOAP (over HTTP)
to be defined as requested in the IANA considerations of
this document.
and in the IANA Considerations section:
The IANA is requested to assign TCP ports for NETCONF for
SOAP over HTTP and SOAP over BEEP.
Would it be preferred to revise the first paragraph to contain
the following?
The NETCONF SOAP service MUST be offered
over the new standard port for NETCONF over SOAP
Thanks,
Ted.
Thanks,
P.S. http://xml.coverpages.org/draft-presuhn-nmwebdav-01.txt
describes (at a very high level) a network configuration protocol
that doesn't tunnel over HTTP, but instead uses HTTP/WebDAV as an
application protocol. The use of SOAP is not described though, but
you could imagine all of the configuration documents wrapped in a
SOAP envelope. Such a use of HTTP for network configuration would
probably comply with BCP 56.
Mark.
--
Mark Baker. Ottawa, Ontario, CANADA. http://www.markbaker.ca
Coactus; Web-inspired integration strategies http://www.coactus.com
--
to unsubscribe send a message to netconf-request@ops.ietf.org with
the word 'unsubscribe' in a single line as the message text body.
archive: <http://ops.ietf.org/lists/netconf/>
--
to unsubscribe send a message to netconf-request@ops.ietf.org with
the word 'unsubscribe' in a single line as the message text body.
archive: <http://ops.ietf.org/lists/netconf/>