OPINION

IN PRAISE OF OPEN ARCHITECTURE

By F. A. Putnam

At the Industrial Computing  Roundtable (reported last issue) I was very surprised to find myself practically the only supporter of Open Architecture for industrial computing. This was a shock, as at our company, we believe strongly in Open Architecture. In  fact, I'm convinced our company wouldn't even exist without it!

What is Open Architecture? It is a system that has openly published interfaces that permit products from multiple vendors to work together as part of a system. Users want it. Why? It allows them to mix and match products from different vendors, which increases their options and lowers their costs.

Open Architecture tends to go hand in hand with standards. A product can have an open architecture without having anything to do with any standard. As soon as that product exists, however, it is a candidate for becoming a de facto standard. Standards can either be defined by committee, or be adopted de facto.  The process of standard setting by committee is slow and political, since every vendor wants their system adopted as the standard. Some believe the best way to get standards is de facto. How standards happen is less important than  their existence, and what we do with them.

Some important Open Architecture standards for industrial computing are 1) the IBM-compatible PC (personal computer) standard, 2) standard networks like Ethernet, and MAP, 3) 110 interface standards like IEEE-488, and 4) software connectivity standards POSLX.

Benefits of open architecture Let there be no mistake about it, I am a strong proponent of Open Architecture. Why? Well, I have  personally seen the benefits of the PC standard as it has unfolded over the past ten years. This computing system is a de facto standard. It was introduced by IBM with a strong commitment to Open Architecture. IBM published the specifications of the PC's interconnect bus, printed the source code to the ROM BIOS firmware in every manual, and documented very clearly just exactly what parts of the system were considered part of the standard, and were never going to be changed. For example, every PC-compatible has an Intel 8253 counter-time chip at an address 40H, and has a clock input to this chip of 1.19318 MHz. The definition of this standard allowed the PC to be adapted to literally millions of vertical applications. This let the PC become a high volume product, which drove the price down and fueled the growth of what is now an entire industry.

For one who lived in pre-PC days, the contrast is  striking. In those days, our company was selling its own computer system. We were a proprietary system vendor. Our hardware only ran our own software, and our software only ran on our own hardware. We had to

 create everything. I admit it — we weren't very successful with this approach. The total potential market for our software was limited to the size of our own hardware installed based.  When the PC standard arrived, we converted to a software-only vendor, and now we can leverage off everything that happens in the PC market. Companies like ours are growing rapidly, and are helping to bring to market open industrial  computing systems of extremely high functionality, quality, and reliability at price levels that are shockingly low compared to proprietary systems.

Users against vendors? - Is Open Architecture a good thing for all parties? No. It's bad for the proprietary system vendors who have a vested interest in the old "account control" practices of locking users into buying every system component from them. These vendors are always going to fight Open Architecture tooth and nail. This was clearly evident at the Industrial Computing Roundtable. These people will tell you that open systems are bad because they stifle innovation. Don't  believe it —there is plenty of room to innovate within a standard. They will tell you that open systems are bad because they lead to systems being sold as commodities, with severe competition from overseas "clone" vendors. Don't believe this either. That's their way of expressing their extreme discomfort that they can no longer sell their proprietary systems at the high price levels they are used to, prices  which are consistent with their very low volumes of production.

In today's market, savvy users are voting for Open Architecture with their purchasing dollars — in droves. The result of this voting is an overwhelming victory for the open systems vendors (e.g., IBM, Compaq, Sun). Vendors who were formerly oriented towards closed systems (HP, Data General, Wang, NCR, etc.) are discarding these old ways  and embracing openness wholesale. Why? It's the only way they can survive. Victory for the users

!ABOUT THE AUTHOR

Dr. Frederick A. Putnam is a Ph.D. and M.A. in chemical engineering and  B.A. in physics. He was graduated from Carnegie-Mellon, Case Western Reserve, and Dartmouth. He is president of LAB TECH and a widely published author.

home | about | products | services | support | seminars | library | links