service/resource identifier on the wire?
In the past I've advocated that we need an AbstractName equivalent as a service/resource identifier on the wire, i.e. present in the soap message. During the last conference call it was mentioned that at the ws-naming F2F it was discussed that this was not needed. Could someone explain what the arguments were? What the discussion was about exactly? (maybe the f2f-minutes would be nice too ... hint, hint ;-) ) Thanks, Frank. -- Frank Siebenlist franks@mcs.anl.gov The Globus Alliance - Argonne National Laboratory
Dear All: Frank sent a nice email a while ago making the case for this. I think it's really unfortunate that there was no response. This lack of response to substantative proposals and questions (except, sometimes, for a response that "this was discussed and was not needed") seems to be standard operating procedure for WS-Naming and OGSA-BES. I don't think it is the right way to work if we want consensus and adoption. Ian, At 08:48 PM 11/8/2005 -0800, Frank Siebenlist wrote:
In the past I've advocated that we need an AbstractName equivalent as a service/resource identifier on the wire, i.e. present in the soap message.
During the last conference call it was mentioned that at the ws-naming F2F it was discussed that this was not needed. Could someone explain what the arguments were? What the discussion was about exactly?
(maybe the f2f-minutes would be nice too ... hint, hint ;-) )
Thanks, Frank.
-- Frank Siebenlist franks@mcs.anl.gov The Globus Alliance - Argonne National Laboratory
_______________________________________________________________ Ian Foster www.mcs.anl.gov/~foster Math & Computer Science Div. Dept of Computer Science Argonne National Laboratory The University of Chicago Argonne, IL 60439, U.S.A. Chicago, IL 60637, U.S.A. Tel: 630 252 4619 Fax: 630 252 1997 Globus Alliance, www.globus.org
Frank, et al, Let me make sure I understand your question. When you say "on the wire" do you mean "must be present in the soap header"? If so, then the discussion was to change the document to state that the endpoint receiving the message MUST NOT assume that the header contains the abstract name. The "endpoint" in this case refers to an endpoint that is being "named". Thus, it cannot be used for dispatch. I cannot recall who suggested this (it might have been Tom, I cannot recall though). The motivation was that because not all elements of the WS-Address are guaranteed to be included in the header by all tooling. If not, what do you mean by "on the wire"? Andrew -----Original Message----- From: owner-ogsa-naming-wg@ggf.org [mailto:owner-ogsa-naming-wg@ggf.org] On Behalf Of Frank Siebenlist Sent: Tuesday, November 08, 2005 11:49 PM To: ogsa-wg@ggf.org; ogsa-naming-wg@ggf.org Subject: [ogsa-naming-wg] service/resource identifier on the wire? In the past I've advocated that we need an AbstractName equivalent as a service/resource identifier on the wire, i.e. present in the soap message. During the last conference call it was mentioned that at the ws-naming F2F it was discussed that this was not needed. Could someone explain what the arguments were? What the discussion was about exactly? (maybe the f2f-minutes would be nice too ... hint, hint ;-) ) Thanks, Frank. -- Frank Siebenlist franks@mcs.anl.gov The Globus Alliance - Argonne National Laboratory
Andrew Grimshaw wrote:
Frank, et al, Let me make sure I understand your question.
When you say "on the wire" do you mean "must be present in the soap header"?
Yes.
If so, then the discussion was to change the document to state that the endpoint receiving the message MUST NOT assume that the header contains the abstract name. The "endpoint" in this case refers to an endpoint that is being "named". Thus, it cannot be used for dispatch.
I cannot recall who suggested this (it might have been Tom, I cannot recall though). The motivation was that because not all elements of the WS-Address are guaranteed to be included in the header by all tooling.
I'm not sure if I understand what the discussion was related to. It is clear that the current AbstractName element will not end-up in the soap-header. This was the reason for me to suggest the use of the Address as an alternative to the AbstractName and by using a profile to add the right uniqueness properties to the Address such that it can be used as an identifier. Are there issues with the tooling and the address value, and how it ens up in the soap header? And even more important: does the current ws-naming working group believes that having a service/resource identifier that is also present in the message is important enough to standardize its use? -Frank.
-----Original Message----- From: owner-ogsa-naming-wg@ggf.org [mailto:owner-ogsa-naming-wg@ggf.org] On Behalf Of Frank Siebenlist Sent: Tuesday, November 08, 2005 11:49 PM To: ogsa-wg@ggf.org; ogsa-naming-wg@ggf.org Subject: [ogsa-naming-wg] service/resource identifier on the wire?
In the past I've advocated that we need an AbstractName equivalent as a service/resource identifier on the wire, i.e. present in the soap message.
During the last conference call it was mentioned that at the ws-naming F2F it was discussed that this was not needed. Could someone explain what the arguments were? What the discussion was about exactly?
(maybe the f2f-minutes would be nice too ... hint, hint ;-) )
Thanks, Frank.
-- Frank Siebenlist franks@mcs.anl.gov The Globus Alliance - Argonne National Laboratory
Hi Frank, I think the following email thread is what you looking for. # To: ogsa-naming-wg@xxxxxxx # Subject: [ogsa-naming-wg] WS-Names and WS-Addressing WSDL Binding # From: Maguire_Tom@xxxxxxx # Date: Fri, 7 Oct 2005 12:41:09 -0400 # Cc: tuecke@xxxxxxxxx
So this says to me that the <wsa:address> is essentially equivalent (or at least could be) to an abstract name.
http://www-unix.gridforum.org/mail_archive/ogsa-naming-wg/2005/10/msg00000.h... Hope it helps, ---- Hiro Kishimoto Frank Siebenlist wrote:
In the past I've advocated that we need an AbstractName equivalent as a service/resource identifier on the wire, i.e. present in the soap message.
During the last conference call it was mentioned that at the ws-naming F2F it was discussed that this was not needed. Could someone explain what the arguments were? What the discussion was about exactly?
(maybe the f2f-minutes would be nice too ... hint, hint ;-) )
Thanks, Frank.
Hi Hiro, Hiro Kishimoto wrote:
... I think the following email thread is what you looking for.
Thanks but no, I know that email well and believe that Tom shows a good path forward that addresses the requirements that I articulated. Probably it is the email thread that Andrew is looking for, though ;-) It would be good to get some statement of direction whether this is an accepted approach for inclusion in the ws-naming spec or how we could get to such a decision... Regards, Frank.
# To: ogsa-naming-wg@xxxxxxx # Subject: [ogsa-naming-wg] WS-Names and WS-Addressing WSDL Binding # From: Maguire_Tom@xxxxxxx # Date: Fri, 7 Oct 2005 12:41:09 -0400 # Cc: tuecke@xxxxxxxxx
So this says to me that the <wsa:address> is essentially equivalent (or at least could be) to an abstract name.
http://www-unix.gridforum.org/mail_archive/ogsa-naming-wg/2005/10/msg00000.h...
Hope it helps, ---- Hiro Kishimoto
Frank Siebenlist wrote:
In the past I've advocated that we need an AbstractName equivalent as a service/resource identifier on the wire, i.e. present in the soap message.
During the last conference call it was mentioned that at the ws-naming F2F it was discussed that this was not needed. Could someone explain what the arguments were? What the discussion was about exactly?
(maybe the f2f-minutes would be nice too ... hint, hint ;-) )
Thanks, Frank.
-- Frank Siebenlist franks@mcs.anl.gov The Globus Alliance - Argonne National Laboratory
participants (4)
-
Andrew Grimshaw -
Frank Siebenlist -
Hiro Kishimoto -
Ian Foster