to Cyreenik book index

The Diamond in the Rough

Ray Discovers the LAN

Less than two and a half years after he joined Boschert, Ray was scouting for new entrepreneurial opportunities. For the 12 years after he left GE he had devoted most of his energies to building up businesses that other people had started. Now he wanted to try his hand at founding his own company.

In 1987 Ray recalled the circumstances that ultimately led him out of Boschert and into Novell.

I had actually started a business attempting to do some phases of what we are doing here at Novell. My plan was to develop to a typical $50-million to $100-million manufacturing environment [producing] a computer-based system with certain levels of fault tolerance[Footnote 1] that would allow me to run 24 hours a day for $250,000 as a system, and I was aimed at doing just that kind of business.

The reason that I had a compelling interest in doing that was that I had run three other companies doing someplace in the $50-million level of business, and to run the business I had to run computer systems which cost about $1 million to maintain any level of security around running the factory without losing time. So I would typically buy two minicomputers, [DEC] VAXes as it turns out, and put them in hot standby so that if one failed the other one could be switched over as quickly as possible—and it usually took not less than two hours—so that I could continue to run the business.

So in my spirit of entrepreneurship at that time, I decided I was not going to run another business that had to pay a million dollars, so I funded and started a company called Reliable Data Systems that was aimed at using a particular technology in a fault tolerant configuration that would allow us to do just that.

And we were headed in that direction when the folks who had financed Novell asked me to come over and take a look at it. And when I saw it, I saw a better way to do it. I saw a short-term and a long-term existence in the business that would provide exactly the same thing that I had started with a different technology in another company. So after a few months, it was clear to me that I had to stop that other business and come over and run this one, because it offered a better opportunity for everybody.

Ray saw something in Novell Data Systems that had not been in Jack Davis’ business plan when Jack approached Ray two years earlier: The local area network. The printers, terminals, and computers that Jack had been interested in making, Ray had dismissed as low-margin commodity items. But the LAN intrigued him. Here was a product with a future.

Ray looked at the LAN and saw a business computer system that cost far less than minicomputer systems. Because of its simplicity and low cost a business could buy two of them and achieve fault tolerance. By buying a LAN a customer could preserve his investment in his existing stand-alone PCs, and as a business grew it could simply add more low-cost PCs to the network.

By distributing processing power over a network of intelligent PCs, the LAN overcame the limitations to growth that minicomputer users eventually ran into. In LANs each user has their own “smart” PC which processes information. The more PCs you add to the network, the more powerful the network becomes.

With minicomputers and disk sharing, by contrast, each user taps into the central processing unit of the host computer via “dumb” terminals, and the more users you hook up to the central processor the less performance each user sees. As a growing business added users to its minicomputer system, the computer would take longer to process data and eventually the system would have to be replaced. Ray saw that the LAN could overcome the “planned obsolescence” of the minicomputer.

Another appealing aspect of the LAN was that it harnessed the prevailing technological trends. By 1982 it was clear that there would be a rapid proliferation of personal computers in the 1980s and beyond. Every person or business that owned more than one PC would be a prospect for a LAN.

Finally, the heart of the LAN was a high-margin software product, not a low-margin hardware item, and this appealed to Ray. As software the network operating system was inexpensive to manufacture, difficult for competitors to steal or emulate, and easily revised or customized. Properly handled ShareNet could become for LANs what Digital Research’s CP/M had become for microcomputer operating systems: An industry standard.

At the end of November 1982, Ray stopped by the Novell Data Systems booth at Comdex in Las Vegas—a small booth with a wall of the new IBM PCs networked with a wall of older CP/M-based micros. He introduced himself to the attractive blonde in the booth, Judy, and said, “I’m interested in your company.”

No Founders

There was one other interesting difference between Novell and the other companies that Ray had worked with. Novell had already cast out its founders—there was no Goshorn, Zschau, or Boschert left at Novell to later contest his success.

This would be Ray’s company from the start.

Footnote 1: As we have seen, fault tolerance is the term used to describe the ability of a computer system to continue operating even in the event of a component failure, usually achieved by creating parallel or back-up systems.

to Cyreenik book index