OPC or Mimosa for XML?

R

Thread Starter

roger Irwin

I am ever more convinced that the future of plant wide supervison and industrial control lies in the hands of XML, and it would seem that most
would agree with this, certainly when you ask anybody about it they say "thats the way we are heading".

Yet whilst other industry sectors, toolmakers and OS vendors (yep, even MS) are coming up with standard schemas, Automation seems a bit left
behind.

OPC are supposedly working on it, but I cannot find any official offerings. Indescretions say that they aiming to use XML to transport
thier existing DCOM objects so that it is upwards compatible. One naturally assumes that this is uninformed speculation, I certainaly hope so. DCOM is heavily typed whilst the versatility of XML is due to its abstraction and free form nature. COM over XML sounds like fitting a
Ferrari with tractor tyres, and if the bloated nature of OPC systems reminds you of the Michelin man, this approach should come up like a Sumo wrestler;-). So what are OPC REALLY doing with XML? Anybody know?

The only organisation that seems to be active on XML for industrial automation is Mimosa (http://www.mimosa.org). This is a much smaller
group than OPC, and they do not make much noise, but they appear to be a busy and well informed bunch. Well, at least they have allready defined
and released a working XML standard, and defined conformance tests, and indeed certified a couple of industrial XML products. Do they have a
future?

Is anybody else woking on XML schemas for IA?
 
H

Hullsiek, William

> Roger Irwin:
> Is anybody else woking on XML schemas for IA?

I have seen references to a XML standard being developed by the OAG (Object Automation Group)
that is to be used for interfacing to 'Factory Floor Systems', i.e., MES. I would like to have ERP systems send the data in XML to a Biztalk server, that build the work orders for MES.

However, I do not have several thousand dollars to join OAG. (Buying standards from ISA is a bargain, compared to some of these other industry groups).

Bill Hullsiek
MES Software Engineer
 
A

Alex Pavloff

> I am ever more convinced that the future of plant wide supervison and
> industrial control lies in the hands of XML, and it would seem that most
> would agree with this, certinly when you ask anybody about it they say
> "thats the way we are heading".

Am I the only one here who can't figure out how this makes sense? Lets says that some foundation puts together an XML dictionary that everyone can agree on (ok, now quite laughing... this is only hypothetical). Great. Now what?
How do you send the XML documents around? Do you expect every single device on an industrial network to parse the XML? XML parsing, while svelte compared to most applications on a desktop, does take up (remembering some figures here, hopefully not out of my butt) on an embedded Linux system around ~600KB of flash.

> Yet whilst other industry sectors, toolmakers and OS vendors
> (yep, even MS) are coming up with standard schemas, Automation seems a bit
left
> behind.

Yeah, what else is new?

> OPC are supposedly working on it, but I cannot find any official
> offerings. Indescretions say that they aiming to use XML to transport
> thier existing DCOM objects so that it is upwards compatible.

Ok. Now I'm wondering how the heck this make sense. I know how COM works. I understand the concept behind DCOM, and its based off of Sun RPC, and how XML applies to this, I have absolutely no clue. They seem to me to be two
in two completely different technical areas. If someone can tell me how XML is going to be used to transport DCOM objects (one doesn't transfer DCOM objects anyway... they sit in one place while you make RPC calls).

> One natrually assumes that this is uninformed speculation, I
> certinaly hope so. DCOM is heavily typed whilst the versatility of XML is
due to it's
> abstraction and free form nature. COM over XML sounds like fitting a
> Ferrari with tractor tyres, and if the bloated nature of OPC systems
> reminds you of the Michelen man, this approach should come up like a
> Sumo wrestler;-). So what are OPC REALLY doing with XML? Anybody know?

COM is not DCOM. This is a common misconception. COM over XML makes absolutely no sense to me.
 
M

Matthew da Silva

There is no doubt about this being the way of the future not only in control but also in design and engineering (and, hence, purchasing and ecommerce). The only way we will get a standard schema is by doing it ourselves. Leaving it to MS or anyone else is a sure way to get some new type of part solution that has a lot of 'embedded intent' to be compatible with their own interests. Also, since it has a scope that is large and ubiquitous like Fieldbus, if the user community would get involved in greater numbers from the early stages, then the user community would possibly support its product with greater enthusiasm.

The other thing that I think is that, the only way to do this is to do it online. No other way can accomodate the depth and breadth of the
requirement, and the time aspect. It needn't be supervised by ISA (but if it is, that's cool too) since many of ISA's active members would take part in it.

The model is not mimosa, which seems to me to have a little less specific gravity than would possibly be required to pursue an XML schema for
automation. Instead I would suggest looking at the rosettanet initiative which is a daughter organization of commerce Net.

Matthew Yamatake Tokyo

-----Original Message-----
On Behalf Of roger Irwin

Is anybody else woking on XML schemas for IA?
 
M

Michael Griffin

At 00:26 04/11/00 +0100, roger Irwin wrote:
<clip>
>The only organisation that seems to be active on XML for industrial
>automation is Mimosa (http://www.mimosa.org). This is a much smaller
>group than OPC, and they do not make much noise, but they appear to be a
>busy and well informed bunch. Well, at least they have allready defined
>and released a working XML standard, and defined conformance tests, and
>indeed certified a couple of industrial XML products. Do they have a
>future?

You or someone else has mentioned MIMOSA before. They are working in a very specialised area of instrumentation. I imagine what we would see in industry in general is not some single overall organisation, but rather a lot of specialised ones.

The one really wide spread general purpose application I can think of is production counts and efficiencies. This is something everyone running a business would like to know, and there would be a very large demand for this if the information was easy to get at.


>Is anybody else woking on XML schemas for IA?

Why don't you start your own organisation? Write a standard, wave it around, puff it up with lots of press releases and le voila, its official. That's what everyone in the computer industry does, isn't it?

Seriously though, do you have something specific in mind, or are you just curious about this subject? I believe this isn't the first time you have written about it. If you have some ideas about how to get a *usefull* standard developed, I would be interested in hearing about them.
Too many of the standards organisations involved in software and communications in this industry seem to be either of the "mountain has
groaned and brought forth a mouse" sort, or they are secretive incestuous little groups who don't listen to anyone except themselves (and no I'm not
referring to the ISA in either case). This is why so many standards efforts have been a failure.

I have a real bad feeling though that this area is going to end up like the fieldbus mess, only much worse.


**********************
Michael Griffin
London, Ont. Canada
[email protected]
**********************
 
F

Frank Iwanitz

Roger,
you are right OPC Foundation is doing something on combining XML and OPC technology.
The working group has met already several times.
Probably a first specification will be issued spring next year. You can download the mission statement from the OPC foundation web site.
The working group discusses two things:
- a framework, i.e. how to use, transmit, secure ... XML documents
- XML schema for the different specs, starting with Data Acces. Probably there will not be a direct mapping of DCOM calls on XML schema.
The integration of OPC technology in internet applications is one of the goals, therefor schema will contain aset of requests. The integration of non-MS-systems is another aim. I hope this gives you some more insights in what OPC is doing.
Regards,
Frank
 
R

Ralph Mackiewicz

> I am ever more convinced that the future of plant wide supervison and
> industrial control lies in the hands of XML, and it would seem that
> most would agree with this, certinly when you ask anybody about it
> they say "thats the way we are heading".
>
> Yet whilst other industry sectors, toolmakers and OS vendors (yep,
> even MS) are coming up with standard schemas, Automation seems a bit
> left behind.

Not really. Check out the ISA SP95 standards effort. They are developing standard schemas for exchange of data between controls and management
systems (to put it simply). Go to:

http://www.isa.org/sc/committee/1,1293,145,00.html

> OPC are supposedly working on it, but I cannot find any official
> offerings. Indescretions say that they aiming to use XML to transport
> thier existing DCOM objects so that it is upwards compatible. One
> natrually assumes that this is uninformed speculation, I certinaly hope
> so. DCOM is heavily typed whilst the versatility of XML is due to it's
> abstraction and free form nature....snip...snip...So what are OPC
> REALLY doing with XML? Anybody know?

As far as I can tell, OPC is really looking at using SOAP instead of DCOM. SOAP is a way for remote procedure calls to be encoded into XML for
transfer over a network. SOAP will be non-proprietary and portable to other platforms versus DCOM which is proprietary. Remember, OPC is an API.

> > Is anybody else woking on XML schemas for IA?
>
> I have seen references to a XML standard being developed by the OAG
> (Object Automation Group) that is to be used for interfacing to
> 'Factory Floor Systems', i.e., MES. I would like to have ERP systems
> send the data in XML to a Biztalk server, that build the work orders
> for MES.
>
> However, I do not have several thousand dollars to join OAG.
> (Buying standards from ISA is a bargain, compared to some of these
> other industry groups).

Try the other OAG: Open Applications Group. Much of the OAG's work is available for free download after you register. These are more business
related documents. However, it is important to see how the OAG architecture works before moving on to industry specific stuff if you intend to leverage mainstream technology for your XML activities. Go to:
http://www.openapplications.org

Regards,
Ralph Mackiewicz
SISCO, Inc.
 
For the only information I've seen that ties it all together in a particular way, check out the architecture document on the UPnP website
http://www.upnp.org/resources.htm

XML is the encoding of the information over the wire. SOAP is the way RPC commands and responses are encapsulated in the HTTP protocol, I think.

As you allude, that's the easy part. Now lets get busy on that dictionary (schema).

Best,
B.O. Nov. 7, 2000
--
Robert Old, System Architecture, [email protected]
Siemens Building Technologies, Inc., HVAC Division
1000 Deerfield Pkwy., Buffalo Grove, IL 60089-4513 USA
Phone: +1(847)941-5623, Fax: +1(847)419-2401
 
A

Alex Pavloff

> For the only information I've seen that ties it all together
> in a particular way, check out the architecture document on the UPnP
website
> http://www.upnp.org/resources.htm

Thanks for the info. I can see how it works together now.

> XML is the encoding of the information over the wire. SOAP
> is the way RPC commands and responses are encapsulated in the HTTP
protocol, I think.

Wow. We've just started to get devices with ethernet ports on them that pretty much replicate what the serial ports did previously. With luck,
we've got TCP/IP on them, and given even more luck, they're actually robust enough to use. Now we're supposed to use SOAP on HTTP using XML to talk to devices? And you want them for how much? And you want them when? You can make all the standards schemas you want, but I have very serious doubts about using the desktop and server level communication schemes on embedded systems.

Maybe someone who's put a system together can tell me roughly what it took (cost, size, etc etc) to put together a system running a TCP/IP stack, web server and XML parser together, because that's what it looks like those of
us pushing toward a "standard" communication scheme are trying to make us use. Call me crazy, but that looks way too expensive. Not every company can charge an arm and a leg for communications modules for a "standard"
communication scheme and get away with it.
 
M

Michael Griffin

At 09:42 06/11/00 -0800, Alex Pavloff wrote:
<clip>
>Am I the only one here who can't figure out how this makes sense? Lets says
>that some foundation puts together an XML dictionary that everyone can agree
>on (ok, now quite laughing... this is only hypothetical). Great. Now what?
>How do you send the XML documents around? Do you expect every single device
>on an industrial network to parse the XML? XML parsing, while svelte
>compared to most applications on a desktop, does take up (remembering some
>figures here, hopefully not out of my butt) on an embedded Linux system
>around ~600KB of flash.
<clip>
It sounds like a lot until you realise that Siemens offers a web and e-mail module for their S7300/400 PLCs (and possibly other companies may offer something similar). Why not an XML module?

I wouldn't expect XML to be supported by field devices. It would be useful though to be able to connect machines or cells to the rest of the company in a standard manner and extract various data to improve efficiency, quality, productivity, etc.
Right now all you can really do in most cases is some form of custom register mapping. This makes every application unique and adds to the work (and cost) of connecting machines into larger plant systems. It also makes interfacing every OEM machine a custom job if it can be done at all.

You can pay for 600K of flash pretty quickly if you can eliminate (or minimise) system integration time in an application like this. This of course would only work if there is a readily available universally accepted standard that is independent of the actual controller hardware and operating system. I see this standard as being a far bigger hurdle than the hardware issues.


**********************
Michael Griffin
London, Ont. Canada
[email protected]
**********************
 
R
Michael Griffin wrote:

> You or someone else has mentioned MIMOSA before.

Me, some time ago, but things did not get clearer.

> >Is anybody else woking on XML schemas for IA?
>
> Why don't you start your own organisation? Write a standard, wave it
> around, puff it up with lots of press releases and le voila, its official.
> That's what everyone in the computer industry does, isn't it?

Automation industry! When I do something like this it is with tongue in
cheek;-)

> Seriously though, do you have something specific in mind, or are you
> just curious about this subject? I believe this isn't the first time you
> have written about it. If you have some ideas about how to get a *usefull*
> standard developed, I would be interested in hearing about them.

XML has great potential for remote supervision and control of field devices and I am convinced it will play a large part in IA future, so I
want to get into it now. As for 'how to create/manage a standards group, the most succseful computing standards group of all time (arguably the **only** one everbody has conformed to over a long time scale (thirty odd years)) is the IETF. The very name of their standards 'Request for Comment' reeks openness. A peep at their archives wil reveal an enourmous quantity of standards, many unused. That about sums it up, let everyone contribute schemas/implementations to an archive (paid for by
banners/ad sponsors), and let the most popular thrive! Survival of the fittest. It's the Darwinian way..........

> I have a real bad feeling though that this area is going to end up
> like the fieldbus mess, only much worse.

Quite. Deja Vu.

I suppose part of my interest in XML is my hope that it will break us out of this benign egoistic mess.
 
W

Wallinius Mattias

Depending on what OS you use. For example VxWorks ships with the IP stack and webserver. XML parsers are normally free of charge under GPL. If you are concerned about cost use Linux, Apache and XML Apache. Tomcat is also interesting as a full scale JSP server. You will also get a very "scalable" and performant solution. "Scalable" meaning from small embedded computers to large servers. It will require a lot of investments in knowledge from your
organisation. For real-time behaviour there is KURT, RT-Linux or RTAI. Remember that TCP/IP, HTTP and XML never will do real-time.
/Mattias
 
I hear you there. The assertion so far is a further addition to the mnemonic plague, SCP:

http://www.microsoft.com/HOMENET/scp/default.htm

If we can make it happen, think how useful a universal communication service can be. More research to come.

Best,
B.O. Nov. 8, 2000
--
Robert Old, System Architecture, [email protected]
Siemens Building Technologies, Inc., HVAC Division
1000 Deerfield Pkwy., Buffalo Grove, IL 60089-4513 USA
Phone: +1(847)941-5623, Fax: +1(847)419-2401


-----Original Message-----
From: Alex Pavloff [mailto:[email protected]]

&lt;snip>
devices? And you want them for how much? And you want them when? You can
make all the standards schemas you want, but I have very serious doubts
about using the desktop and server level communication schemes on embedded
systems.

&lt;snip>
 
Top