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

Re: Roy's ApacheCon presentation

From: Mark Baker <distobj@acm.org>
Date: Wed, 20 Nov 2002 13:48:38 -0500
To: Ugo Corda <UCorda@SeeBeyond.com>
Cc: www-ws-arch@w3.org
Message-ID: <20021120134838.H21537@www.markbaker.ca>

On Wed, Nov 20, 2002 at 09:23:17AM -0800, Ugo Corda wrote:
> I looked at presentation slides 4 and 5 in particular: "EAI - the hard way" and "EAI - the Web way". 
> The first picture looks completely misleading to me. No EAI product approaches the problem that way (one-to-one interactions), and in fact the acronym EAI has become synonym with the idea that you do not do integration that way.

Sure, ok, you're O(Nlog(N)) as a best case, and O(N^2) as a worst case
(thanks Miles).  Still, would you agree that that's significantly worse
than O(N)?

> An example of an EAI architecture, that is both based on current EAI architectural direction (called by some "Enterprise Service Bus") and on Web services and XML technologies, can be found at [1] (Universal Application Network, or UAN - I suggest you download the white paper).
> 
> As Anne mentioned before, reusability of interfaces is crucial in such a context. The UAN architecture actually goes even beyond that. Relying on reusable interfaces and a standardized choreography language (hopefully something will get standardized soon ...), the range of reusability can be pushed to the level of entire business processes.

We're getting off track here.

I'm trying to emphasize that, for the purposes of integration, it's
better to have less interfaces than more.  You'd agree with that, right?
I'm also saying that if you could define an interface which could be
used to wrap all information systems, then this would be a wonderful,
glorious thing, since you can't do any better than having one interface
(though that doesn't preclude there being more than one interface that
can wrap all information systems).  Do you agree with that?

I suggest that the interface defined by HTTP is such an interface.
To Anne, GET/PUT/POST/DELETE plus mandatory extensions and URIs, are
complete; anything you can do with custom interfaces, I can do with this
one interface.  Plus my cost of integration will be substantially lower
than yours, and the network effects I generate will be MUCH stronger.

Sigh.  I thought this was a simple point!

MB
-- 
Mark Baker.  Ottawa, Ontario, CANADA.   http://www.markbaker.ca

   Will distribute objects for food
Received on Wednesday, 20 November 2002 13:45:05 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 3 July 2007 12:25:10 GMT