Re: "Reliable" web services for Next Big Thing? (was RE: Agenda for 5 December WSA telcon)

Hi Anne,

On Thu, Dec 05, 2002 at 02:09:50PM -0500, Anne Thomas Manes wrote:
> WSA is protocol independent.
> 
> I'd like to be able to support identical reliability metrics regardless of
> the underlying transfer protocol.

And I'd like to be the King of Spain! 8-)

There are two different ways to use SOAP (architecturally), and each has
its own way to be protocol independant.  I agree that protocol
independance has its advantages, but in a very different way than you
mean, I think.  I'm sure we don't need to get into this now, but I've
written a little about it in the past.  See;

http://www.markbaker.ca/blog/2002/10/30#2002-10-soap-definition
http://www.markbaker.ca/2001/07/SoapUses/

So the impact to reliability is, sure, some aspects of reliability can
be done in a protocol independant manner; message identity, specifying
an expected reliability QoS with a mandatory extension, maybe some
other things that aren't occuring to me right now.  But the methods of
the application protocol matter too, and can help with reliability in
many situations.  For example, if you did an HTTP PUT, but the
connection died before you heard back, you could invoke GET to find
out if your message got there or not.  You can't do that with SMTP.
Like I say, different application protocols have different coordination
semantics, and simply do things differently.  But many can be *extended*
in the same way, which is what SOAP is really good at.

Hmm, I'm sending a lot of messages again.  Am I helping?

MB
-- 
Mark Baker.   Ottawa, Ontario, CANADA.        http://www.markbaker.ca
Web architecture consulting, technical reports, evaluation & analysis

Received on Thursday, 5 December 2002 15:03:32 UTC