Networking: Emergence 1979-1981
LANs and DataPBXs
7.17 Ungermann-Bass: Xerox, Broadband and Needing a Chip
With each passing month of 1980, UB management, regardless of role, felt the stress to find customers. Charlie Bass discovered the first significant opportunity when, over a dinner with David Liddle in San Francisco, he learned Xerox needed a product to interconnect their Star product to asynchronous terminals and peripherals of other vendors. Bass pitched Liddle on using UB’s NIU-1s that UB would OEM to Xerox. Intrigued, Liddle asked if UB would be willing to conform their version of XNS to Xerox’s and upgrade to the 10 megabit per second Ethernet. Without hesitation Bass committed to doing so, knowing that being able to claim interoperability with Xerox’s products gave UB valuable credibility. Before the end of the year a handshake sealed the agreement with specifics and timetables to be negotiated. Davidson remembers:
"Only then did they [Xerox] realize that they had forgotten to connect in any foreign equipment, and then they looked around for how they would do that quickly, and they found our terminal servers to fit the bill if we would A) adopt the ten megabit standard, because we had built four megabits on our own, and B) change the protocols to be compatible with what Xerox was actually using, which we were delighted to do."
Aside from the handshake with Xerox, for a product that would not be available until well into 1981, UB found being first to market little consolation. It seemed to matter little to investors whose enthusiasm for networking investments let UB close their second round of venture capital financing in November, raising $4.972 million on a nearly eight times step-up in company value to $19 million. For fiscal, and calendar, year 1980, UB had sales of $436,000 with a loss of $1.2 million
To convert the Xerox handshake into formal contracts and then manage the on-going OEM relationship, management courted Judy Estrin of Zilog. Following the exodus of engineers to UB in 1979, Estrin had been promoted to engineering manager for networking, and Carrico had become business unit manager. At the June 1980 National Computer Conference trade show, Zilog had introduced their Z-Net LAN, a scaled-down Ethernet made possible by the dissemination of information out of PARC and the social structure existing between the employees of PARC and Zilog. Zilog’s future, however, turned increasingly on decisions made by Exxon Office Systems and networking seemed to interest no one. Estrin remembers:
"We spent an infinite amount of time trying to deal with Exxon Office Systems -- trying to get them to use the Z-Net technology. It was just, politically, very difficult.”
By year-end 1980, Estrin had had enough and, assured by Ungermann that he had learned the importance of focus from his Zilog experience, she joined UB on February 1, reporting to Jordan and responsible for the Xerox account. Estrin remembers:
“One of the things I felt at Zilog is they were doing too many things, trying to be too much, and I even remember sitting with Ralph before I accepted the offer and saying: "You did this at Zilog. Are you going to do this again?" "Absolutely, positively, not. We're going to focus, we're going to do this," and I got to UB, and they had their first product out, and they were going in a million directions.”
Exploration of directions even included a conversation with Evans of Micom, their dataPBX nemesis. Bass remembers
"I remember meeting with Roger Evans. We were looking at ways -- they had used Z-80s, so we knew each other from the Zilog days, and we were sizing each other up, looking for ways to cooperate --the wonderful thing about this business is you can sit down with your arch enemy and explore common interests, and I think we were trying to see how these technologies might converge, how we might do something together. It didn't get anywhere, but we certainly had the conversations."
And after Sytek introduced their broadband LocalNet 20 in February 1981, UB management found they had more than dataPBXs to contend with. For it did not take long for the low cost and futuristic appeal of broadband to handle voice and video (not to say anything about the hatred that existed between UB and Sytek) to prompt UB management to justify broadband products as consistent with their vendor independent strategy. To jump-start development, Bass suggested they recruit Gregory Hopkins of MITRE, an acknowledged expert in broadband and holder of patents regarding CSMA/CD over broadband. In June, Hopkins joined UB starting an office Burlington, MA. Jordan remembers:
“We started this broadband engineering operation in Burlington. It worked for me. I don't remember why. Looking back on it, it's probably a great way to start a rift.”
Engineering also had to be gearing up to introduce new, lower cost products once Ethernet chips were available from Intel. Ungermann knew the risks of waiting but there seemed little choice given the expense of doing their own silicon. Then luck knocked on their door and an alternative appeared. Bass remembers learning about Silicon Compilers, a new start-up:
“We were approached by Kleiner Perkins, who was then an investor, and they had this company that had a silicon compiler. They came and they wanted to test their technology on something hard. So we said: "Well, we've got something hard for you. We'd like to have an Ethernet chip." So they come in and they started running equations and talking to Alan Goodrich, who was our hardware designer. And we learned a lot about their tools and what it's going to take to build one of these.”
With everything going UB’s way, in July 1981 they announced their fully compliant 10 megabit Ethernet NUI executing fully compliant XNS. Success seemed assured.