RE: Reliable Messaging: Division of Responsibilities (was RE: RES T, Conversations and Reliability)

As a WG member...

>> *If* we assume that the view that applications must share *some*
>> responsibility for reliability, then I think our task is to add
>> as many semantics and formalized processes to access and manage
>> these semantics.
>
> Can you elaborate on what you mean by that?  I don't understand.

Thanks for the support and I will try to be more specific. 

Assumptions:
1) basic layered architecture: transport, application
	(Pick what ever words work for you, these are the labels I
	am most comfortable with.)
2) intermediaries may be part of a layer or may represent their a
	new one---I am not debating that at this time.

3) an interesting and significant part of an architecture is to
	+ describe the roles and responsibilities for the layers
	+ describe how they interact

Proposed work items:

1) create strawman for layers (or equivalent)
2) place intermediaries in one (or more) locations in the arch
3) describe reliability as it relates the layers
4) identify roles and responsibilities relative to the layers.
etc.

Clearly, there will be several models of what reliability is
and how to achieve it. SOAP, REST and all the other perspectives
will need to be considered. But at least when we are done it will
be more constructive than simply agreeing to disagree.

Dave H

-----Original Message-----
From: Mark Baker [mailto:distobj@acm.org]
Sent: Thursday, August 08, 2002 12:16 PM
To: Dave Hollander
Cc: www-ws-arch@w3.org
Subject: Re: Reliable Messaging: Division of Responsibilities (was RE:
REST, C onversations and Reliability)


Dave,

On Thu, Aug 08, 2002 at 10:51:08AM -0700, Dave Hollander wrote:
> The REST, Conversations and reliability thread has been
> informative. I am splitting out what I believe to be one of the
> key design issues to come from it in hopes of focusing and
> progressing to a useful architecture document.

Good idea.

> *If* we assume that the view that applications must share *some*
> responsibility for reliability, then I think our task is to add
> as many semantics and formalized processes to access and manage
> these semantics.

Can you elaborate on what you mean by that?  I don't understand.

MB
-- 
Mark Baker, CTO, Idokorro Mobile (formerly Planetfred)
Ottawa, Ontario, CANADA.               distobj@acm.org
http://www.markbaker.ca        http://www.idokorro.com

Received on Thursday, 8 August 2002 17:41:05 UTC