Re: Summary, 9-11 Nov 2004 WS Description WG FTF: two objections

On Fri, 19 Nov 2004 19:56:13 +0600
Sanjiva Weerawarana <sanjiva@watson.ibm.com> wrote:
> So, does this mean that the current in-out pattern can ONLY be
> used with a bi-directional underlying transport? In particular,

I don't see why.  But it suggests that the replyto address ought to
identify the same "node" as the requestor.

When the replyto address is permitted/expected to identify a different
"node," then it's a different MEP.

> if I'm doing SOAP/SMTP instead of SOAP/HTTP, then clearly I have
> to use RFC822 ReplyTo (or the usual From: hackery) to figure out
> where to reply to .. is that correct or incorrect use of the 
> in-out MEP in your view?

In this case, I would say that the correct use of the MEP is only when
the replyto address is the "same node" (FSVO the term) as the "From"
address.  Possibly a "logical" node, but presumably, if the address
belongs to someone else running different software and hostile to my
interests, they're not quite entirely the same node.  That is, "best
practice" holds that for protocols that mandate replyto addressing and
inband correlation, the current in-out MEP should have the same
semantics that we've given it (which are based on HTTP's semantics).  If
you wanna send the response to *some other node*, then the MEP ought to
be a *different MEP* that permits that (as a best practice, mind,
because once you've got the replyto address, there's no policing that
this is what's actually happening).

> > The "generalized" pattern of in-out would have a request from some
> > node N1, a response to some node N2.  Whether N1 and N2 are the same
> > node would be left unspecified.  They MAY be, but are NOT REQUIRED
> > to be.
> 
> Even the SOAP request-response MEP does not require that the
> response be delivered via the same network connection with the

I'm not at all certain why the obsession with network connections, here.

> bits going the other way. IMO we have to reflect reality that
> its very natural for a node to originate a message via one
> "connection" and receive a response via another "connection."

*shrug*  I don't believe that you can represent that as in-out in WSDL
2.0 as it now stands.  I agree it's an important use case, but it
requires some sort of extension to enable it anyway.

> I don't want the use of WS-Addressing to be "hijacking" anything.
> If there's a problem with the status quo let's fix it .. IMO

It currently does, because WSDL 1.1 has a very limited vocabulary of
MEPs.

> IMO this is unnecessary; we should just define "node" logically and
> be done with it ... then messages can be sent and received at
> different host/port combos (or whatever such beasts) which in
> combination is the"node."

Unless they *aren't* a "node," and aren't linked administratively,
corporatively, by blood, marriage, oath, fire, orientation, gender,
species, planet, language, or anything else.  Which is the point. 
Replyto can point at something which is definitely and emphatically
*not* the same node.  We ought to provide a means of saying "this is the
MEP you use then.  When you use the in-out pattern [that is, the current
one], your replyto SHOULD identify the same node."

Amy!
-- 
Amelia A. Lewis
Senior Architect
TIBCO/Extensibility, Inc.
alewis@tibco.com

Received on Friday, 19 November 2004 20:55:33 UTC