# MS 'Monopoly'? was ENGR WinNT Reliability

E

Does this mean that Linux with its anti-competitive prices (0$) is liable for anti-trust motion??? Anand E #### Eminent Hi Curt, the "yada...Yada..." Stuff on ISA magazine made great reading. But time and again all of us in the list seem to be fighting an unknown devil. But is the Devil Microsoft? As far as I know Microsoft is a very recent entrant into Automation (With PC based SCADA and Soft PLC's working on NT or Windows 95/98/2000/XP). In fact are they to blame because some Automation vendors picked this platform up because it was more acceptable to the general public. remember that prior to this there were several Unix based costlier systems around and most of the systems were on propreitary platforms. If Linux becomes more aceptable to the public then these same vendors may bring out products on Linux Platform. This is why Linux must make inroads into the Desktop in a big way. M$ may have done so many wrongs (imagined or true or whatever), but it is popular! And that is the chief reason that many automation vendors have
products based on M$. Secondly development on M$ may be more economical and faster for an Automation vendor than Linux.

Again system features are more important. If somebody creates a system based on M\$ with no thought to reliability issues then who is to be Blamed. I have seen UNIX systems with reliable platforms but improper automation system
implementation. This system had several upgrades on graphical interface, Engineering station and IMS. And it used to hang! You had to reboot and one such reboot caused the plant to trip because of a mistake in configuration of the stop system loop!! I would not blame Unix here.

What the Automation world needs is a reliable automation system. Again Reliability is also dependent on the engineering done while implementing. Cost is also critical but not factor "numero uno". For a HMI, in a well configured system a once in a while blue screen may be acceptable to the user!

Everyone agrees that Linux is great and yet there are no major Linux products in automation systems around. If someone would make a working system on Linux and it succeeds then these discussions will have more relevance.

As against IT, Automation systems are moving out of monopolies on UNIX systems and proprietary OS's to open systems on Microsoft. In fact systems
on Microsoft and the facility of inter and intra communications have reduced the costs in automation systems.

Microsoft may have done some lobbying, created products that facilitated faster development of systems, but this is just marketing strategy. Linux needs to emulate. In addition Linux has the advantage of being able to upgrade some of the older UNIX based systems.

Microsoft is not certainly a charitable institution and neither were the other OS's. Linux is an Exception to this, but until a workable solution comes out.... lets wait!

Anand

E

#### Eminent

The question is can you create a product which more than say 70% of people like and buy and if you have created such a product weather this will become a monopoly.

Anand

C

#### Curt Wuollet

One of the first Linux automation projects I did was a similar case only on a much smaller scale. As rebuilders, WMC gets back a core for nearly every unit they sell. They wanted to automate handling those. A system for checkin I had developed while in IS was expanded to many parallel stations and produced barcodes that were stuck on the packages. This system was running Linux already as we are a UNIX shop. From the checkin stations the boxes were placed on a sorting conveyor that was to disposition the boxes according to barcode. We were brainstorming to design a conveyor when management decided to outsource it. We were somewhat perturbed, but when asked if _we_ could do it in two weeks, we responded that obviously that would be impossible. End of discussion.
They picked a firm that sells Quantum conveyors. An NT based product that they had seen at some show. For months they diddled and putzed and tried to make that system work. Through this period, I was rather glad I was not involved as they were having to fill trailers with cores for checkin and nothing was moving. The last couple weeks, every manager in the plant was sorting missorts and our systems administrator was standing by to boot the NT system on an average of every 20 minutes. I got looks that could kill when I was driven to laughter by the spectacle. After a little while of the most expensive material handling on the planet it got to lawyer time and we got a challange. Can you _fix_ it in two weeks, and if you can, you get a bonus. Order what you need.
There wasn't time to do it entirely with PLC's as the comms suck or with Linux as I didn't have the industrial IO hardware I have now, so we did the logic and fifo on the PLC and the comms, database and display on the Linux box with a cmm interface between them. We beat the deadline and have had half an hour of downtime in two years, due to a devicenet master on the PLC, Fortunately we had a spare. We shipped three pallets of NT equipment, proprietary electronics and assorted crap back to the vendor and did the job with Beckhoff devicenet IO and a small GE 90/30 and, of course the same Linux box that had been there to run the checkin stations. We used their mechanicals which aren't that great but they do work. This is one of the type of wins you savor on many levels. A small stuffed penguin sits atop the console and smiles. It was the end of NT on the factory floor here forever, or at least until they forget.

Regards

cww