Ungermann-Bass

» see ub.com as it lived, on the wayback machine

~ the goodbye ~

Ralph Ungermann had co-founded Zilog with Federico Faggin in 1974, watched Exxon's money arrive, and left at the end of 1978 over what that money wanted. In 1979 he started again in Santa Clara with Charlie Bass, another Zilog man, on $1.5 million of venture funding and a bet that networking was going to be a business by itself rather than an accessory sold by whoever made the mainframe in the corner.

That was not obvious in 1979. Networks came from computer companies, matched to their own machines, and a company with terminals from two vendors had two networks. Ungermann-Bass shipped the Network Interface Unit in July 1980, the first Ethernet product you could buy without buying a computer to go with it, and built Net/One on top of it: a system whose selling point was that it did not care whose equipment you plugged into it.

For about eight years that was the whole story of enterprise networking, and Ungermann-Bass was the biggest independent in it. Then it made the wrong call in one meeting, and after that it was somebody's subsidiary until the last owner switched it off in 1997.

~ the chronicle ~

~ the meeting about routers ~

In 1986 an engineering group inside the company, under John Davidson, built a working IP router prototype and found a customer willing to deploy it. Management turned the product down. The reading upstairs was that customers wanted bridges, that bridges were what the installed base understood, and that a router was a solution to a problem nobody had written a purchase order for yet. Cisco had been founded two years earlier and would spend the next decade taking exactly that market.

Every engineer has sat in some version of that meeting. The thing that gets said is never stupid, it is the sensible summary of what customers have actually asked for, and the honest part is that at the time you usually cannot tell which prototype on the bench is the next decade and which one is a distraction. What makes this one worth writing down is that the company had the router, the customer and the engineer, and still filed it. 1986 was also the year it earned $18,000 of net income on $110.9 million of sales, which is a rounding error dressed as a profit.

~ the nine years as a subsidiary ~

Tandem Computers announced the purchase on 22 February 1988 at $260 million, $12.50 a share in cash plus a debenture. Ninety five percent of the stock was tendered by 18 March and the deal completed on 30 June. Ungermann became a vice president and board member at Tandem, and the company he had built to be independent of computer makers was now owned by one.

It kept selling. Renamed UB Networks, it held revenue somewhere between $350 million and $377 million a year through the mid nineties with a base of ten to twelve thousand corporate customers, which sounds healthy until you look at what those customers were buying. The money was in hubs and bridges for legacy wiring at the exact moment every network in the world was being rebuilt around switches and routers. Fiscal 1996 came in with operating losses estimated as high as $80 million.

~ the actual end ~

Tandem signed a letter of intent with Newbridge Networks on 19 December 1996 and the sale closed on 17 January 1997 at a base price of $118 million, which is less than half of what Tandem had paid nine years earlier for a company that was then much smaller. Newbridge said the right things about the installed base and then did the arithmetic. The GeoSwitch ATM switch stopped shipping in May 1997 after about nine months on the market, leaving around eighty sites with a product that had no future, and the plan on paper was to move a third of the inherited customers onto Newbridge's own line.

What Newbridge wanted was the ten thousand accounts and the four hundred sales people who knew them. The Santa Clara operation was wound down over the following months and most of the staff let go. Ungermann went on to a videoconferencing company and later to investing in China, and died on 2 June 2015. The idea he started with, that the network is its own industry and not a feature of your computer vendor, won so completely that it is hard to explain to anyone now that it was ever a bet.

@}-->--

              )
             (_)
            .-'-.
            |   |
            |   |
            |   |
            |   |
          __|   |__   .-.
       .-'  |   |  `-:   :
      :     `---'     :-'
       `-._       _.-'
           '""""""

« previous  ·  random  ·  next »