New forum topic - Open Control

K

Thread Starter

Ken Crater

Hi all,
People who have been regulars at Control.com and on the Automation List will recall that our principal mission as an organization is to encourage the advent of open control, in the sense of multi-vendor standards-based systems.

We feel that a global peer community of working automation engineers is an important step toward making this vision possible. We are well on the way toward accomplishing this goal, with over 50,000 unique visitors each month, and the world's most active discussion venue for automation topics.

I'm pleased to announce that we are now introducing a new topic to our website and to the companion Automation List, that will be a gathering place for discussions on open control technologies. This topic, Open Control (OPENC), will include discussion on:

- open control architectures and components
- open interfaces, including:
--bus structures (cPCI, PC/104, ISA, PCI, etc.)
--device-level busses (Modbus, Profibus, DeviceNet, etc.)
--networking (TCP/IP and related protocols)
--APIs
--language standards
- open source software
- portable control software
- interoperability issues
- performance and benchmarking of open system components

We hope this new topic makes an important contribution to our industry and to the people who make up the Control.com forum. We're inviting you all to contribute to the discussion, share experiences, and help the controls world become a bit more open.

Regards,
Ken Crater, President
Control.com Inc.
[email protected]
 
B
Hi Ken Crater:

Can you give me your definition of "open control", one that everyone can agree to?

Bob Pawley
250-493-6146
 
Hi Ken,

A significant local symposium for the 'PC-Based Control' was held in Seoul last fall.
The first topic in the meeting was to define the exact meaning of 'PC-Based Control', since it has been called some different names, such as 'PC-Based Control', Open Control, Softlogix, SoftPLC,
PC Control and ...

If this forum focus on 'PC-Based Control',
I think it is timely forum and my company can share our valuable experiences with PC-Based Control applications.

Thanks...
Key

 
B
Hi Key:

Does your company define 'open control' as being based on a PC or virtual hardware, as opposed to other hardware?

Bob Pawley
250-493-6146
 
K
In response to Bob Pawley...
>Can you give me your definition of "open control", one
>that everyone can agree to?

Bob Old posted:
>And the way to World Peace, too, while you're at it, Ken.

Nah, World Peace is too easy (in comparison) :).

Actually, I would not equate "open control" with PC-based control. It is conceivable (and, in fact, has been done) to create a closed control
system using a PC-architected system. All you have to do is use a proprietary bus structure, write closed-source code, and use a proprietary protocol -- result: a closed PC-based system.

I don't hold to any finite criteria as a threshold of "openness", but rather see it as a question of "preponderance of evidence". Peter has already mentioned some of the key attributes of an open system, which I'll elaborate:

- adherence to published, freely-usable standards at key interfaces (bus, protocol, physical form factor, signal levels, etc.).
- commercial availability of meaningful choices from a variety of vendors in an open market, for any unit of functionality.
- unbundled support available at a variety of levels from a selection of providers.

Although not absolutely necessary (in my book) to have an open system, open source software is certainly helpful in meeting the above criteria.

The whole point of open control is to reduce cost and risk, and increase flexibility to respond to change. This can only be accomplished byopening each component choice to multiple suppliers competing in a free market. Instead of being "stuck" with the offerings of a single company, the user gets to choose best-in-class technology from among dozens of companies for each piece of the puzzle.

I think that's why people have been talking about this for 15 years. And before anyone says it's impossible, it's already happened in the PC world. Go to the Dell site and look at the menu choices for any component of one of their systems:

> ATI, RageT 128 Ultra, 32MB, VGA [subtract $109]
> nVidia, Quadro2 EXT, 32MB, VGA Dell Recommended
> nVidia, Quadro2 ProT, 64MB, VGA/DVI [add $311]
> ATI Fire GL2, 64MB,VGA/DVI [add $611]
> ATI, RadeonT VE, 32MB, VGA (dual monitor capable) [subtract $49]
> ATI, RadeonT VE, 32MB, VGA/DVI (dual monitor capable) [subtract $29]

If I've learned anything in 25 years in this field, it's that you need to have an enormous toolbox to handle the diversity of applications that come along even in a single plant. One vendor just can't keep up with the pace of technology in all requisite fields anymore. That's why we thought it was time to make open control a practical reality.

Best Regards,
Ken Crater, President
Control.com, Inc.
[email protected]

World Peace *next* year...
 
B
Thanks for answering Ken.

I wonder what form open control should take.

Is it the Linux model where anyone and everyone who does some work in the plant is able to 'remodel' the control software in any way he or she desires? Where innovative solutions are either freely distributed or remain a plant asset? Where the effort of supporting the plant's software continues to burn out highly skilled workers and retains the $100+ per I/O price tag?

Or should we look for another solution to open control - a solution that is cost and time effective? A method of incorporating proprietary and "free" software in such a way that control solutions can cross various protocols, hardware (both virtual and real) and vendors offerings. Solutions that will assist the user in their implementation cutting costs and opening new business opportunities.

Personally, I have always had a problem with 'free' that has no structure nor boundaries. Everything has to be anchored to something real; and those people developing solutions, real solutions should be rewarded for their creativity.

Bob Pawley
250-493-6146
www.automating-automation.com
 
H

Hullsiek, William

Personally, the definition works for me. It covers the Foxboro situation where they sell you a Sun (which is open) but you only buy the Sun from Foxboro.


From a user-perspective, vendors can (should) add value by:

1. Testing

You probably can 'charge more' if you test your components with other vendors.
Let your customer know they have a choice.

2. Improving manage-ability of the components.

Have a unified frameworks based on CORBA, Java, SNMP for managing multiple vendor control equipment.

3. Different programming tools / IDE for porting

Would be nice to have CTC state logic running on other equipment.

4. Training courses -- Application instead of vendor.

Imagine if we had training courses organized by application instead of by vendor.
It would really be useful, if we could lower the cost of implementation and ownership for our customers.

- bill hullsiek
 
A

Anand krishnan Iyer

I believe that this is probably the best possible definition of what is practical and what the user should want from an open system.
Anand
 
R

Ralph Mackiewicz

> - adherence to published, freely-usable standards at key interfaces
> (bus, protocol, physical form factor, signal levels, etc.).

If the standard costs $100 to purchase from a standards organization does this preclude its use in your definition? Or does "freely- usable" mean no restrictions on use such as royalties?

> - unbundled support available at a variety of levels from a selection
> of providers.

I am curious as to what "unbundled support" means. I have a preconception that it means that if you buy some software that you must also purchase support. Or would the situation where support is simply optional meet this criteria?

> The whole point of open control is to reduce cost and risk, and
> increase flexibility to respond to change.

I think it is important to clarify what is meant by "cost". There are many elements of cost including purchase price, installation, packaging, support, maintenance, modification/flexibility, productivity/throughput, etc. There is a tendency to focus only on purchase price for cost reductions which makes many kinds of useful and necessary investments look impractical. Or, it incorrectly puts the spotlight on something like free software (for instance) as being the source of the cost reduction which results in disappointment and therefore failure.

> This can only be accomplished by opening each component choice to
> multiple suppliers competing in a free market.

...snip...snip...

> And before anyone says it's impossible, it's already happened in the
> PC world.

Nothing is really impossible, but there are significant differences between the PC market and the IA market that need to be considered. The massive investments that took place in the PC market by hundreds, if not thousands, of companies that enables this interoperability was driven by the returns from a market that is a couple of orders of magnitude larger in size. That is why I don't think analogies with the PC industry are all that relevant. But, I don't think that it is impossible for IA either. All that is really required is an energized user base that understands and buys open technology. Everything else will flow naturally from that.

Regards,

Ralph Mackiewicz
SISCO, Inc.
 
K
Ken Crater:
> - adherence to published, freely-usable
> standards at key interfaces(bus, protocol,
> physical form factor, signal levels, etc.).

Ralph Mackiewicz:
If the standard costs $100 to purchase from
a standards organization does this preclude
its use in your definition? Or does "freely-
usable" mean no restrictions on use such as
royalties?

Ken Crater:
The latter. In open-source parlance, this is "free" as in free speech, not free as in "free beer". In other words, a standards org can charge
for the paper the standard is printed on, but the intellectual property content of the standard must be open for free use if it is to meaningfully be called an "open" standard.

Ken:
> - unbundled support available at a variety
> of levels from a selection of providers.

Ralph:
I am curious as to what "unbundled support"
means. I have a preconception that it means
that if you buy some software that you must
also purchase support. Or would the situation
where support is simply optional meet this
criteria?

Ken:
If there is no requirement to pay for support in order to obtain the hardware/software, and the hardware/software is not "opaque", then the
option becomes available to obtain third-party support at the best price/competence levels for your company.

Ken:
> The whole point of open control is to reduce
> cost and risk, and increase flexibility to
> respond to change.

Ralph:
[...] There are many elements of cost including
purchase price, installation, packaging, support,
maintenance, modification/flexibility,
productivity/throughput, etc. [...]

Ken:
True, and it is in the consideration of the full range of these cost elements that open technologies become the most attractive. Openness has its true value in the expansion of options, which can benefit the user throughout the lifecycle of a system. As Ralph points out, the near-term issue revolves around the support infrastructure, which has grown in the computer industry and which we are helping to grow in the controls industry (including on this forum <grin>).

Ken Crater
Control.com Inc.
[email protected]
 
C

Curt Wuollet

Hi Bob

That's an excellent question. As a practical matter we have nearly all closed systems for financial reasons. Lock-in and incompatibility do serve the purpose of maintaining a revenue stream. I think there are other models that would work. Nirvana would be for there to be equal opportunity for everyone to compete for seats at a particular solution table. We Open folks are obviously ready to do that because it would certainly be much easier than having to reverse engineer everything. An all Open world would not neccesarily be the best either as there is valid need to recover investments etc. Unfortunately they run the world at the moment and are highly unlikely to move in a direction that involves risk for them even if it would provide great benefit to the consumer. The only way that I see the situation changing is if Open Solutions gain enough traction to make openness a requirement that shows up on RFQs and requirement documents. Since the major vendors have shown zero willingness to cooperate amongst themselves, I doubt that it is realistic to ever expect them to cooperate with Open Systems providers. That's why as I have said before, Openness tends to be absolute rather than relative and it will be an either or situation until it begins to seriously affect their revenue and they have no other choice.

There is absolutely no doubt that a mix of closed and open and more choice would be better for the consumer. Unfortunately there is no other way to get there than to buy Open Systems until they play ball. This requires that there be Open Choices available. That's what we are trying to do. We have a chance because we don't have
to make big money doing it. As close as I can tell, no one wants choice, but we're betting they will once we make the benefits tangible.

Regards

cww
 
B
Hi Curt:

It would seem there is a bipolar reality in the control world these days.

At the south pole are numerous control software packages. Systems that do the job, but are so rigid in application that everyone finds some fault.

At the north pole is Linux, the ultimate open source software, but without the structure and organization to instil the confidence and acceptance required for mass appeal.

Somehow these apparently disparate systems need to meet at or near the equator with at least some pretence of harmony.

I suggest that the present situation is not sustainable. When these two bipolar systems do harmonize there will be an explosion of change within the industry. Those who prepare for this change will be the only ones who survive.

Bob Pawley
www.automating-automation.com
250-493-6146
 
J
So paraphrasing, let me know if I understand the gist of what you are saying.....

Rockwell, GEFanuc, and the others are completely closed and currently have no intention of untrenching.

You and others that promote the 'completely open' approach are taking up the position that is diametrically opposed, in the hopes of convincing the 'closed systems' providers to meet somewhere in the middle?

Just checking my bearings......

--Joe Jansen
 
B
Joe

I don't know about others, I speak only for myself.

I came into this open control and open source discussion later than most with no preconceived ideas other than what I have lived with for the last 30 some years.

It does seem to me that we as an industry have moved from total proprietary systems, to a point where we aren't absolutely required to purchase hardware and software from the same source.

The other extreme that has come about because of Linux causes me to note that there is a definite movement toward systems being as open as possible. This trend will continue regardless of what individuals think.

What I am suggesting is that a totally open system such as Linux - standing by itself - is to much the opposite extreme to have great appeal. Some boundaries, some anchor is needed so that all concerned have a solid place to stand in order to see what needs to be done or to see what has been done.

Proprietary systems have this solid place upon which everyone, technician to corporate bureaucrat, is comfortable. The fact that this proprietary place is so vast gives me reason to consider a melding of the two extremes.

Bob Pawley
www.automating-automation.com
250-493-6146
 
S
> So paraphrasing, let me know if I understand the gist of what you are saying.....
>
> Rockwell, GEFanuc, and the others are completely closed and currently have no
intention of untrenching.
>

I don't think that you can throw all vendors of proprietary systems into the category of being uncooperative in support of Open Systems technology. Rockwell appears not to support the notion of untrenching. However, GE Fanuc and
Schneider Automation have stated their realization that the industrial automation market will, however slowly, move towards open systems. A recent discussion that I had with an executive at GE Fanuc was encouraging. He stated that the proprietary PLC market has reached the point where PLCs are a commodity. GE Fanuc desires to participate in the open systems movement in
order to raise their controllers from the commodity market to becoming a component of an open system. This will allow them to once again differentiate their product based upon its superior performance as an open system component
for specific applications. With respect to Schneider's efforts in this area, they have product available today which can be used to build an open control system and they have the courage to support an open forum discussion on their company website's home page. Schneider has absolutely no control over the postings to their forum as the forum is moderated and managed by an
independent, third party (Control.com). One will get you a hundred that you will never see that coming from Rockwell.

Steve Harr
Control.com
 
H

Hullsiek, William

Suggested improvement to definition of Open Control

One attribute of an Open Standard is the availability of source-code examples, i.e., done under Gnu or University type of sponsorship, hence the popularity and low cost of TCP/IP.

Many vendors offer the IP protocol suite at a nominal charge, because they can take "Open Source" and port to their hardware at a minimal effort. Which is why *nix systems and IP work so well in a commercial environments.

FieldBus and MAP would have been done deals by now, if they had sponsored a University or other organization to develop the software /firmware "prototypes" and place them in the public domain.

- bill hullsiek
 
C

Curt Wuollet

Wiser words were never spoken.

It really helps to make something ubiquitous if you act like you _want_ people to use it. The attitude and impression I get from the fieldbus consortia involves buying memberships, equipment and licenses or they'll sue me for using it. Completely kills _my_ enthusiasm. I would think they would like for us to popularize their systems with MAT/LPLC. It might be a pretty good investment.

Regards

cww
 
I don't think it is possible to define "OPEN" Control as if it was some standard for a protocol. OPEN is a philosophy that embodies vulnerability and innovation. If you have a product that is OPEN, you are vulnerable to lose your customer to another choice. However, if you are the most innovative and provide your customer a high degree of on-going value, vulnerability becomes an asset. Being vulnerable is the juice that drives small or major corporations to soar to new levels.

In most cases the major industrial players don't see themselves, or the market this way at all. In fact the larger you become (mergers & acquisitions) the less vulnerable you are, the more you control the market. Being huge means you can buyout trade publications to dominate advertising, impress decision makers for large accounts with trinkets and might, for example. The technology of automation is complex. Few people understand the full ramifications of a decision one way or another. Those that do are lower down the corporate ladder. Every situation is different, the factors too numerous to list. But size, capacity, installed base, and the ability to invest thousands on entertainment and sales campaigns are a big part of obtaining large orders, and keeping the lid on OPEN innovators.......

Here is a piece I am working on for our next newsletter aXFlash. I spent 8 years of in systems engineering and a about a year in sales/marketing for the #3 automation vendor in the US. This is a candid view from the inside. (Note - Advisory - It contains some promotional statements about our company). Comments or additions are welcome!


>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>.

Confessions of an Industrial Salesman

Delivering increased value to customers is something rarely discussed inside regional meetings of large industrial vendors. The focus of discussion is usually the volume of orders booked and how to generate more revenue in order to meet the sales targets. The products and sales practices of firms are often designed to maximize revenue "pull" from the customer base.

Proprietary communications, proprietary devices, hardware, and software provide vendors with market protection by being exclusive. They exclude competition from other vendors for expansions, upgrades, spare parts and service. New customers are given very aggressive pricing to secure the facility.

Are these practices "industrial atrocities" and are they short sighted? Do such strategies harm the long-term competitiveness of the very industries they serve?


A seasoned industrial systems salesman chortled "You know in most businesses you give your good customers a break, but in this business once they are locked in you put the screws to them. We low-balled the bid to get the order, now we are trying to get back as much as we can by gouging on the transmitters and spare parts."

Large DCS and PLC supply is a proprietary business - for both hardware and software. Even in today's market that has a healthy supply of inexpensive IT components, the major industrial vendors continue to deliver closed systems
despite claims to the contrary.

After customers purchase a system they find upgrades, expansion, parts and services that are expensive. There isn't much one can do once "locked" in. After millions in software development, training, and hardware there is no easy alternative.

The explosion of information technology, standardized operating systems and networks show promise for change within the industrial community. However, we are seeing very slow movement in terms of inter-operability among devices.

The same is true for networks. There is a large, vendor-biased and fragmented choice for industrial networks, some of which are very expensive to implement. In contrast IT rides on widely used standards, with new technology or enhancements adopted relatively quickly as standard for leaps of functionality.

We are seeing industrial facilities operating in markets that are more challenging than ever. The costs of automation systems and IT are significant in the capital or and maintenance budgets. And automation and IT integration play a key role in how the facility manages operations and how efficient the employees can be.

There is no doubt that the use of IT technology in a comprehensive way for in automation lowers the initial and yearly costs of our systems. Also to replace hardware, in particular PLC or DCS processors and their distributed programs with software executing on IT servers and clients is another way to deliver increased value to automation users. The cost of maintaining automation software on the IT platform is far less than hardware in the industrial environment.

It's clear that solution providers must deliver their absolute best effort to clients for new and on-going business. Above all else, top value or price for functionality to system users should be the goal. These key automation systems and their benefit are vital to the long-term health of end-user companies, and in turn the providers.

(Product X ) is designed from need of the user or customer. It's core development came from customers who were after increased value and reduced complexity. (Product X) replaces hardware with software, optimally uses IT technology, and delivers high-performance, flexibility and reliability.

Cheers,

Paul Jager
CEO
www.mnrcan.com
 
G

Greg Schiller

I would agree. People who use technology only want one thing from it. Does it work. The next thing they want is if it is broke where can I get a replacment. Forklifts do amazing things for our economy.

Totally free, open system can fill niches but no one dedicates real R and D dollars to making something they may never own in the future. Imagine it from the venture capitalist's viewpoint.

I would like to take your money to make this very cool niche product that I will then give the rights away to the public. How am I going to pay you back? I'm not. I can't. No one has bought my idea from me.

Venture caps and Corporations with money hate backing up this stuff. It goes against everthing they were taught about in business school. The only way this gets any legs is if it is tagged to hardware sales. But then isn't that why we started this whole mess anyway?

What suffers the most are longterm big payoff ideas that have this type of ending. It took our government persuing a distributed information model to come up with TCPIP. That is the only invester I know that would underwrite large projects for the benifit of all.

Don't take me to far away though. I'm sick too of paying the vender manytimes over the original cost of the product. Membership fees...,
maintenence fees..., buy the hardware and by the way you need to buy our software to make our hardware run...

Open has been most effective when it is used as communication standard. MODBUS RTU/TCPIP, TCPIP, RS232, RS485, USB, Ethernet Etc. Published and then practiced.

Greg Schiller
[email protected]
www.automatica.biz
 
Top