Chapter 12 - Networking: Market Order: LANs 1983-1986
The management of Ungermann-Bass faced the same challenges as 3Com going into 1983. And more. For Ungermann-Bass had adopted a vendor-independent strategy that encompassed more than Ethernet. Even so, Ethernet competition, the announcements of Xerox and Berkeley, and how to react to the personal computer phenomenon added to management’s woes and opportunities.
1983 began with the formal acquisition of Amdax, the source of the token ring technology needed to satisfy their IBM contract. Charlie Bass remembers:
Amdax was really three companies. It was a modem manufacturer in Long Beach, a token ring outfit in Boca Raton, and a software group in southern California. So we shut down southern California. We turned the Long Beach manufacturing group into our broadband manufacturing, and I took over the engineering group, fired the head of that group, and handed them the proposal we had given IBM, and said: “OK guys, this is the job. This is what we have to do.
As the year unfolded, UB felt the competitive pressure from the new entrant Bridge Communications. Ralph Ungermann remembers:
When we started the company we chose the Z-80 as our architecture because of our experience with the Z-80 and the fact that the Z-80 had the best communication capability. That turned out to be our Achilles Heel, because Zilog wasn’t successful, and therefore we ended up having to change architectures to Intel, and that’s how Bridge Communications got started. They recognized that we had a big transition, and they jumped on a 16-bit processor and went around and said: ‘See, they’re the pioneers, but we’ve got the real advanced technology and higher performance and everything,’ and they got a start by getting a jump on us on that second generation.”
That it was Bridge Communications, and not some unfamiliar company, invoked past emotions. Again Ungermann:
We were up here fighting this broad-based battle with this broad product line against IBM and DEC, and then there was the way Bridge got started, really, it caused some real animosity, and there was some tremendous hatred towards Bridge.
UB not only faced the growing head-to-head competition with Bridge Communications but also the lack of a solution to the problems of interconnecting personal computers. Jim Jordan, vice president of sales and marketing, remembers:
The PC started to be a big deal and it started to get very evident to everybody that if you’re going to be in networking, you really had to get into PC networking as part of what you’re doing. Maybe that’s the main strategy, maybe not, but it certainly had to encompass that. I was going crazy, because we were starting to get hurt in the sales area, because we really didn’t really have any network installations for PCs, and it was becoming obvious that they were becoming ubiquitous.
Three guys really ran the company: Charlie, Ralph and I. I wanted to get into it, and I wanted to get into it very badly, and also with a solution similar to the kind of solution that 3Com had, which was as low a cost network interface card for the PC as you could design; no real intelligence on it, and keep the price down as much as you can. That was what I wanted to do. Charlie also thought we should get into the PC networking business, but he said you shouldn’t do it that way. You should build this 186 based board with a whole bunch of memory, and you should off-load all the networking overhead onto this board; take it off the PC, put it on the board, run all your own protocols on it, and cost isn’t really a major issue: typical goddamned engineer approach to the problem. And Ralph didn’t want to get into the business at all. Ralph thought that Ungermann-Bass should not be in the PC networking business at all. The three of us rented a hotel room and spent all day in it arguing about this and trying to work out internal company issues, but this was the main topic. Ralph didn’t want to do it, period. And I’m saying: How can you be a general purpose networking company if you don’t have a PC solution?
So Ralph went to Europe, and Charlie and I got us in the business while he was gone. We absolutely did. We started designing one of these things, and made a commitment with TI. They had, I think, signed a letter of intent with 3Com, and we gave them a technical pitch and said: “We’re going to get into the business. Here’s why our solution is better than 3Com’s, because we’re going to put some memory in and blah, blah, blah…’ and that was when TI was really getting into the PC business, so they called up 3Com and cancelled and signed a deal with us. We took it away from 3Com. We had an OEM deal with TI, and Ralph comes back from Europe and went crazy. It’s the best thing that ever happened to the goddamned company.
Bass remembers the problem of implementation:
In the midst of this, I’m trying to figure out: ‘Can you even get an Intel 186 microprocessor? What’s it going to cost?’ This was at a time when they were on allocation. So I go to Bill Davidow [vice president of Intel], got down on my hands and knees, and said: ‘Please sell us 186’s.’ Bill just played me like a flute. ‘Well, you don’t know, how many do you really think you can sell?’ And this was coincident with 20 companies in the personal computer business – PC business – all claiming 20 percent market share, and forecasting to Intel how many they needed to maintain their 20 percent market share. ‘I need 1,000, and I’ll pay anything, Bill, as long as I can meet TI’s cost, which I couldn’t.’
Investment bankers did not see these internal dramas. They saw the leading company in networking, a market of unlimited potential driven by the growing use of computers. On June 23, 1983, UB became a public company, raising $28.5 million of cash with a valuation of $288 million – roughly half the value of Micom! Far cry from the start-up begging for investors three years earlier.
Then Bass, in one of his frequent trips to Raleigh to manage the IBM token ring-to-broadband product, learned from Dan Warmenhoven that Roger Smith, President of General Motors, has been pressuring John F. Akers, President of IBM, to implement a protocol for LANs being developed by General Motors called MAP for Manufacturing Automation Protocol:
Here I am going to Raleigh regularly, talking to Dan and he starts telling me these stories about Roger Smith calling Akers and telling him: “You’re going to do MAP, or we’re tossing you out, because DEC’s going to do MAP, and if you guys don’t do MAP, you’re not going to do business with General Motors.” And I’m sitting there saying: “Holy shit! This is important. I should pay attention. This is going to happen.” So I start telling Ralph we got to do MAP – that IBM is going to do it, and General Motors is serious about this, Boeing is serious about this. This is really going to happen. So I had a product manager working for me do a business plan, to do a MAP business unit, and we did everything from budgets to site analysis. We were trying to figure out ‘should it be in Austin or Raleigh or where should it be?’ We, at this point, had such a distributed company already, it wasn’t a problem making it more distributed. So we do this business plan and Ralph and Jordan, they’re a little skeptical. I pretty much had Ralph convinced that it made sense. I wanted to do it organically, meaning we would finance it modestly until it started generating some revenues and we could grow the business.
UB finished their fiscal year with sales of $25.4 million and profits of $1.9 million. Ungermann and Jordan, however, with vastly larger visions for the company, wanted to grow much faster. Undoubtedly the conversations they had been having with Bass had had an effect, for as Jordan remembers:
Ralph and I thought the token bus or MAP market was going to be large. I think everybody did at the time. This was early 1984. We decided we wanted to get into it. Concord Data Systems was the only player in it at the time. We felt pretty strongly that their strategy wasn’t that good, because it was very box oriented, and we thought it was a board business. Concord was selling boxes and that was not what you needed in the factory. And they only operated at five megabits, they weren’t ten megabits per second.
The problem was funding the R&D. We wanted to get into it, but here was a whole other development effort, a whole other level of technology, a whole other level of expense. We couldn’t support it from current operations for it would drive our profits down. So we said: “How can we do this?” Well, we could do a R&D partnership. That was one way we could do it. We could do – how about a joint venture company? We both really favored the joint venture company for a couple of reasons. One, we could put the technology in, they would put in the money, but the thing that really made it appealing was if we could do it with the right partner, the thought was that it would lend a lot of credibility. If you go to General Motors and say: “We’re Ungermann-Bass. We’re doing $100 million in sales. You ought to put your whole Saturn plant on our neat technology.” Not too persuasive an argument. But if you go there and you say:” We’ve got this company and GE owns half of it, and GE is standing behind it, you ought to put your Saturn plant on this technology,” it makes a lot better argument.
So we put the proposal together and we went to the Board of Directors. We had a list of three or four companies. GE was the first one on the list. We said: “We’re going to propose to them that we start this company. We want $6 million from them, we’ll give them 40% of the company, we’ll take 60% – keep 60% at Ungermann-Bass. We’ll put in the technology, we’ll manage it, we’ll come out with MAP products.” The board said: “Great,” but they kind of laughed and said: “Good Luck. I’m sure GE’s going to give you $6 million for 40% of this make-believe company,” right? We went to GE in June of ‘84.
During this same period, the spring of 1984, UB signed a $15 million OEM contract with Codex. The terms of the contract gave Codex a warrant to purchase 654,000 shares of UB common stock at an average exercise price of $15.30. (Signifying a very important contract to both firms.)
Managing all of the technology choices and implementations of their vendor-independent LAN strategy kept UB management in turmoil, whether due to: competition, the decision of Xerox to withhold higher-layers of XNS, personal computers or IBM’s contract changes.
Competition with Bridge Communications forced UB to upgrade their products to 16/32 bit microprocessors: if had not been Bridge Communications, it would have been some other firm or simply the need to provide more product functionality. UB began shipping Network Interface Units (NIUs) using the Intel 80186 and Motorola M68000 beginning that third quarter.
The decision of Xerox to not release the higher-layers of the XNS protocol did not have an immediate impact on UB. However, by mid-1984, a decision to convert to TCP/IP or the emerging OSI standard was required. Bass, head of engineering and responsible for the final decision, was very familiar with TCP/IP and had been giving talks on the coming importance of OSI. But maybe most importantly, he had been hearing of how GM had been putting pressure on IBM to support the MAP/OSI protocol stack. Bass opines:
Well, the confusing factor was OSI. I mean – and here’s where I made a huge judgment mistake in gauging the potential importance of OSI. It was one of the biggest flaws in my crystal ball ever. And that’s why I think it’s one of the things that delayed TCP/IP and Ungermann-Bass.
Ungermann, who was intimately involved in pitching the MAP project to GE, also recalls:
We made a huge bet on OSI.
The new 16-bit NIUs conformed to the bottom five layers of the OSI protocol.8
Meanwhile the efforts to develop a personal computer product encountered problems: problems both technical and political. At the root laid the clash between Bass and Jordan as how best to design a PC product. John Davidson remembers:
There were many interactions between marketing and engineering that would probably help paint the picture of the relationship between Jim and Charlie. Jim Jordan and his marketing people would come to a meeting and say: “Well, look, we need this set of features,” and Charlie and his engineering team would say: “Ok, we’ll get back to you and tell you what we can do.” And the next meeting they’d come back and the schedule was enormous and couldn’t meet the needs of the marketing guys for timing, so they’d say: “Well, then, could you do this?” And they’d go away again and come back with another long schedule. This, I think, began to be a problem for the company, certainly a problem for the individuals. One of the questions had been: “Could you put your software inside the PC so that it would run with the dumb card?” And the answer came back: “Yes, but it will take us a year.”
Finally in late 1984, a year after signing their contract with TI, UB introduced its PC product: the Net/One Personal Connection.
In the midst of these technical turbulences, IBM decided to cancel its contract with UB. Bass remembers:
We then suffered from IBM management turnover. In 18 months, we went through three generations of management in Raleigh, and each time they would come in, they would reevaluate, they would re-spec, and affect the project dramatically. The first generation was Murray Bolt. The second generation was Dan Warmenhoven, and the third generation shut it down. They looked and, based on the cost projections, delivery times, they said: “Discontinue.’ This was really my swan song with Ungermann-Bass.
On October 10, 1984, in a big press announcement, UB and GE announced a joint venture: Industrial Networking Inc. (INI). UB contributed technology and management for 60% of the voting stock and GE contributed $2 million in cash and $4 million in notes for 40% of the voting stock. Losses in the joint venture were to be absorbed by GE. Jordan remembers:
We went to GE in June, signed the papers in October, cashed the check. Did the whole deal in about four months.
Before the end of the year, Ungermann elevated both Jordan and Bass to the positions of Executive Vice President. Only Jordan ended up with all the responsibilities and Bass with none. Bass soon resigned, although continuing as Vice Chairman of the Board.
UB finished another record fiscal year. Sales grew 106% to $52 million and profits rose to $6.8 million up from only $.5 million in 1983. Codex accounted for 12% of sales and INI 6%. UB seemed to be clicking in all gears As Ungermann would say in the Annual Report dated February 27, 1985: “1985 promises to be most rewarding for Ungermann-Bass.”
“Ungermann-Bass Net/One,” Datapro Research, May 1985, C11-882-101