Into the ongoing fieldbus wars enters Ethernet

P

Thread Starter

Paul H. Gusciora

From: Manufacturing.Net [mailto:[email protected]manufacturing.net]
Sent: Friday, May 19, 2000 9:26 PM
To:
Subject: Manufacturing.Net Newsletter Issue 183

Email Newsletter from Manufacturing.Net -- http://www.manufacturing.net

Instrumentation and Control

Into the ongoing fieldbus wars enters Ethernet, which is now being used to
connect controllers with remote I/O modules as a fieldbus. Get the scoop at: http://www.manufacturing.net/dc/183/051900w.htm



_______________________________________________
LinuxPLC mailing list
[email protected]
http://linuxplc.org/mailman/listinfo/linuxplc
 
C

Curt Wuollet

Yes, it'a a shame, our good friend ethernet will be corrupted, abused and have its good open nature perverted completely. Even as we speak, the
same bunch of greedy fools who have given us the fieldbus wars are racing to plop a hundred different incompatible proprietary protocols on top of the most successful open protocols on the planet, the internet suite of protocols. A year from now, everything will run on ethernet and none of it will interoperate. And the clueless will wonder why ethernet doesn't seem any better than any other FB.

They just don't get it, these greedy fools, ethernet, TCP/IP, UDP, and the others became absolutely ubiquitous standards on the basis of sharing and cooperation. The best that they can hope for is a tiny piece of the market that they fragmented. It's a shame that nothing and no one can dissuade them from repeating failed strategies from the past. It's an enormous waste of resources and will serve mostly to propel people towards open standards.

I think there is a surprise coming as they try to integrate with the rest of the ethernet/TCP/IP speaking world. Proprietary is a dirty work in
the IS domain already. Even Microsoft, the epitome of the embrace, extend, destroy philosophy speaks TCP/IP and is learning that it's
perversion and decommoditization of open protocols can backfire.

The only thing missing is an Open Source, GPL'ed Industrial Ethernet protocol. Perhaps one of the protagonists will finally look at what is to be gained from the creation of another tower of Babel and take the alternate, popular, goodwill strategy of opening its protocol and becoming the good guy, with the recognition and gratitude that would ensue. They might even glimpse that that is a far larger opportunity than to be an also ran in a fragmented and crowded market. After all,
Deming's definition of insanity is when you keep doing things the same way and expect the results to be different. And perhaps the chance to
actually gain a majority market share might overcome the NIH mindset.
But, sadly, I expect less.

Regards

Curt Wuollet

_______________________________________________
LinuxPLC mailing list
[email protected]
http://linuxplc.org/mailman/listinfo/linuxplc
 
H

Harald Albrecht

Curt sighed at...
> They just don't get it, these greedy fools, ethernet, TCP/IP, UDP, and the
> others became absolutely ubiquitous standards on the basis of sharing
> and cooperation. The best that they can hope for is a tiny piece of the
> market that they fragmented. It's a shame that nothing and no one can
> dissuade them from repeating failed strategies from the past. It's an
> enormous waste of resources and will serve mostly to propel people
> towards open standards.

Every layer above 4 is outlawed by definition (anyone for 5 and 6?). I've even seen FiRo professionals claiming to adhere to OPC when
speaking of 1131 engineering of their systems ... as if OPC would specify any engineering API. It seems to me as if it's common to abuse standards like a popular sport. Layer 7 is good for everything and as both marketing and phb's don't understand what they are talking about, the situation is unlikely to change anytime soon.

> The only thing missing is an Open Source, GPL'ed Industrial Ethernet
> protocol. Perhaps one of the protagonists will finally look at what is to
> be gained from the creation of another tower of Babel and take the
> alternate, popular, goodwill strategy of opening it's protocol and
> becoming the good guy, with the recognition and gratitude that would
> ensue.

Just look at all the global players in the field and count the many (legacy) systems they have. Remember the "Totaly Integrated Automation" slogan? Well... The current situation is that every large company has its own zigurat (generalization class of babel class towers). So how should they even come together?

At which layer(s) should the Industrial Ethernet Protocol (IEP) be specified? Should it replace TCP or even IP and what characteristics should it provide (like QOS)? And then there's still a long way in the automation industry to go to bring IEP into wide operation -- against all the big players, so we will surely going to have a nice time with the all the philistines. Eventually, this should not discourage anybody. I hope we don't have to wait another forty years to occupy the promised Land of Automation and settle.

Regards,
Harald


--
Harald Albrecht
Chair of Process Control Engineering
Aachen University of Technology
Turmstrasse 46, D-52064 Aachen, Germany
Tel.: +49 241 80-7703, Fax: +49 241 8888-238
email: [email protected]-aachen.de

_______________________________________________
LinuxPLC mailing list
[email protected]
http://linuxplc.org/mailman/listinfo/linuxplc
 
C

Curt Wuollet

Harald Albrecht wrote:

> Every layer above 4 is outlawed by definition (anyone for 5 and 6?).
> I've even seen FiRo professionals claiming to adhere to OPC when
> speaking of 1131 engineering of their systems ... as if OPC would
> specify any engineering API. It seems to me as if it's common to
> abuse standards like a popular sport. Layer 7 is good for everything
> and as both marketing and phb's don't understand what they are
> talking about, the situation is unlikely to change anytime soon.
>
> > The only thing missing is an Open Source, GPL'ed Industrial Ethernet
> > protocol. Perhaps one of the protagonists will finally look at what is to
> > be gained from the creation of another tower of Babel and take the
> > alternate, popular, goodwill strategy of opening it's protocol and
> > becoming the good guy, with the recognition and gratitude that would
> > ensue.
>
> Just look at all the global players in the field and count the many
> (legacy) systems they have. Remember the "Totaly Integrated
> Automation" slogan? Well... The current situation is that every
> large company has its own zigurat (generalization class of babel
> class towers). So how should they even come together?

Only when they see that a common protocol is in their best interest, and this requires a great deal less myopia than is prevalent in the industry. This will require a clean break from the past, revolutionary rather than evolutionary.

> At which layer(s) should the Industrial Ethernet Protocol (IEP) be
> specified? Should it replace TCP or even IP and what
> characteristics should it provide (like QOS)? And then there's still a
> long way in the automation industry to go to bring IEP into wide
> operation -- against all the big players, so we will surely going to
> have a nice time with the all the philistines. Eventually, this should
> not discourage anybody. I hope we don't have to wait another fourty
> years to occupy the promised Land of Automation and settle.

My own opinion is that they should follow in the lead of Modbus/TCP and go with an ecapsulation approach. This allows TCP/IP to do what it is good at and builds on the standards. Then it requires nothing special in the way of protocol stacks and will work with commodity hardware, routers, etc. Even the worst possible protocol that interoperates and solves the major headaches they have created would quickly become a de facto standard. Practically anything would be better than the exixting fragmentation and duplication
of effort.

_______________________________________________
LinuxPLC mailing list
[email protected]
http://linuxplc.org/mailman/listinfo/linuxplc
 
Top