W3C home > Mailing lists > Public > www-ws-arch@w3.org > July 2002

Re: [RTF] AC019 proposal to WSA WG

From: Mark Baker <distobj@acm.org>
Date: Wed, 10 Jul 2002 11:26:21 -0400
To: "Cutler, Roger (RogerCutler)" <RogerCutler@ChevronTexaco.com>
Cc: www-ws-arch@w3.org
Message-ID: <20020710112621.K3708@www.markbaker.ca>

On Wed, Jul 10, 2002 at 07:42:36AM -0700, Cutler, Roger (RogerCutler) wrote:
> understanding of what has happened.  So "reliable messaging" does not mean
> exactly that, right?

No, because the end point might be down for longer than any requested
maximum bound on message latency.  So a rejection could be sent by the
queue holding the message.  Sorry, I was oversimplifying.

But typically, reliable messaging solutions provide QoS such as;

- once and only once
- ordering (different types; causal, weak, etc..)

Also, some require that the recipient be notified of a failed reception,
which is effectively requiring end-to-end delivery of information.

None of these are suitable features of a large scale solution to
reliability, IMO.

> Would some of this problem go away if we were more in agreement about what
> "reliable messaging" means?

Perhaps, but I doubt it.  Any use of "reliable messaging" I've seen
assumes that its job is to make the cruel and nasty network look (to
varying degrees) less cruel and nasty.  It is this premise that I
believe is fundamentally flawed.

So if there's a definition of reliable messaging out there that isn't
trying to hide the fact that you're on a network, then I'm all ears. 8-)

MB
-- 
Mark Baker, CTO, Idokorro Mobile (formerly Planetfred)
Ottawa, Ontario, CANADA.               distobj@acm.org
http://www.markbaker.ca        http://www.idokorro.com
Received on Wednesday, 10 July 2002 11:15:04 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 3 July 2007 12:25:01 GMT