High Performance Workstation—Winner takes all

(1986, age 39)

Sun Microsystems was founded in 1982 based on Andy Bechtolsheim’s hardware design and Bill Joy’s software and networking code. The company had growing pains immediately faced with the challenge of producing a commercially viable product. By the start of 1984 with a lot of dedicated work by people too numerous to mention, Sun finally had a workstation on which they could build a company. But by that time they had one more thing: competition.

Apollo Computers was started two years before Sun. By 1986, both companies had done well in the network workstation business, but Apollo was doing better. In late 1986 Apollo had hit $100 million in quarterly sales for the first time. They had twice the market share of Sun Microsystems and were growing quickly.

Something had to change, and it did. The government wanted to standardize on a computer workstation that could be used across many agencies. They put out a bid for an initial $287 million order of workstations to all the vendors, including Apollo, Sun, Masscomp and others. Sun management recognized this for what it really was: a winner take all showdown. Whoever won this first government bid for the “High Performance Workstation” (HPW) would have an approved workstation that any other government agency could order. The $287 million was only a fraction of the value of this bid.

Sun needed someone to head up the technical team that would convince the government that the Sun workstation was what they wanted. I had pulled together teams like this before both at Sun and at previous companies. Clearances weren’t required for this project, that is, they couldn’t tell us anything classified for the bid. However the fact that I did have a military clearances and a thermonuclear “Q” clearance made it easier for me to relate to this project and the people with whom I would inevitably interact.1

I was told to go to McLean, Virginia, just outside Washington, D.C., to set up an engineering team and do whatever it took to win this contract. I could spend any money I needed to spend and pull anyone from the company to come and work on the HPW project. I had been with Sun in the early “bet the company” days and I knew exactly who I needed to make this happen.

There were two parts to what we had to do. The first was a Compliance Demo. In it, we had to show that the Sun workstation could do specific things that the government wanted in a workstation.

The second part of our effort was known as the Flash Demo. In it, we had to show the workstation in a complete scenario that would highlight the workstation’s capabilities.

I wasn’t sure who the actual customer within the government was, but because of where we were physically located, I guessed it was either the NSA or the CIA. I interviewed a few retired people who I suspected could give me some insights into what the unknown customer might be looking for, especially in a Flash demo. The acronym “NSA,” by the way, referred to “No Such Agency” and that’s how we had to refer to it.

Our team started working on the Compliance and Flash demos. The Flash demo would take the most creativity. We had to take all the amazing things the Sun workstation could do and bake it into a seamless, compelling demo. It also had to relate to what I thought the unspecified customer wanted to do with the workstation.2

Scott at HPW

Scott McNealy visited us where we were putting together the HPW demos. I explained what we intended to do and how we would use the workstation. I will never forget how Scott, the CEO of the company, looked at me in disbelief and said “Can our workstations really do that?!” As you see me explaining in the photo, yes, they could and they would.

We needed a plan and I set out to put the entire flash demo on one sheet of paper. As the project manager I knew if it didn’t fit one one sheet of paper then it was too complicated. We needed one page to keep us focused and to maintain perspective. Here is the page we kept taped to the wall.

team leaders

My team leaders were Steve Ferry and Betsy MacLean. With them were a handful of other technical people and support staff. Altogether, there were perhaps 12 people living and working round the clock at our offices in McLean in July and August of 1986.

Under such pressure, and knowing how much was at stake, one might expect there to be internal problems, disagreements, people upset, etc. That didn’t happen. There were clear goals and contingency plans. There were hard completion dates and a two month PERT chart to get us there. Everyone saw the plan and believed that if we hit each mark, we would be successful at the end. If I had gone in with just a “we need to do this by the end” approach, we would not have been successful. It took detailed planning: “do this today” or even “do this in the next three hours.” It all had to fit together, they had to believe it would work and commit to it. It did fit together and the team found a way to hit every milestone along the way. By early September we had given our all. But was that going to be enough?

What a ride it had been! Two months of the most incredible pressure on the whole team. Rarely have I been a part of anything like it. I was proud and amazed at all we had accomplished. I wrote an email moments after the evaluations were completed. It became known at Sun as my “decompression email.” Here it is:

Date: Sat, 6 Sep 86 00:11:05 EDT
From: sundclroger (Roger Frank)
To: nsa@sun.uucp
Subject: NSA LTD

The two day NSA compliance LTD demo is over. Minutes ago the end of the second long day of evaluations was completed. Some of you will be seeing a detailed matrix of how we were rated on each point, and I won’t keep you in suspense by not telling you right off that Sun scored very highly throughout the LTD. I thought that the sunlnsa team might like a parting snapshot from a more personal perspective as well.

We spent two days demonstrating everything we bid in our proposal response to the NSA procurement. Each line item, one after the other, we were asked to substantiate. Those of you who have seen the proposal know it is over two hundred pages long. In the two days allowed, to get maximum points our job was to have a credible demo available and ready to show quickly. There was much to do, and preparation was the key to success.

So how did we do? Just hours ago I announced to the Sun and SDC management that my LTD team had taken all their best shots. We had done all that we could do, and we would wait for the evaluation. Patiently we waited while the NSA evaluation team worked behind closed doors. Then the door opened and we were asked to come in. We sat down around one table: NSA, a few representatives from our prime contractor, and the Sun team.

In a very matter-of-fact manner, the head of the evaluation team started down the list of bid items, one by one. Each was read in order; none were skipped. As each bid item was described, I would recall to myself “That’s the one Keith (DeJager) found a creative solution that finally worked” or “That’s the one Betsy (MacLean) wouldn’t give up on until it was indisputable”. After each item was read and described in a few words, there was a pause while the NSA spokesman scanned across his worksheet to see the results that were recorded. In reality it was only a second or so before he gave the rating. But it seemed much much longer. What’s he going to say? Did the evaluator see enough? Did they believe what they saw? Did we show the right thing? Well? Well…?

“Check.”

And then he would go on to the next line item requirement… read it…delay, and I’d die a little death once again in anticipation…

“Check.”

Another check! Then another. At first I was elated. As each check was announced, I knew that that piece was done. Any concern I might have had could be forgotten now. But wait a minute. “Check.” Is that all? Did they have any idea how hard we worked for some of these? Did we stay up all night, or did we jump through some other fiery hoop just to hear this monotonic “Check.”? On some of them I almost expected an “Amazing” or most anything that would recognize how much work had gone into each line item.

As the time rolled on - and the review lasted hours - the meaning of the nearly unbroken stream of “check…check…” apprasials started to sink in. First, reviewing this all at once made me appreciate even more the amount of work had that been accomplished. Our success was extraordinary and against the odds. We started late; the task was gargantuan in magnitude; there were obstacles at every turn.

Some were external, like the time when the team was starting our “second shift” (which some of you call “night”). As we dug into our tasks, we noticed a sudden stillness and quiet. The air conditioning had been programmed to shut off at 7:30 PM to save a few dollars. We looked at each other and began to type a little faster. Each of a dozen Sun workstations seemed to hum a little louder as the room temperature steadily rose. We worked until the first disks started to fail, and that proved to be one of the only breaks we took during that phase of the project.

Some obstacles were internal. Sun as a company and several people including myself learned a lot about how to prepare and execute a bid as large and complex as this one for the NSA high performance workstation. Our baptism was one of fire. I was lucky enough to have the team of individuals I did have to work with me on the live test demo: people who were bright technically, who remained motivated and focused week after week, and who appreciated the value of teamwork and a proper attitude to get the job done.

But another meaning of that series of “check…check…” winning responses also became clear. That this was over. Nearly the entire summer, each week more intense than the one before, all building up to that moment, now passing into history not with the ticking of seconds on a clock but the calling of the incessant “check”s. Soon, the last check was announced and any areas needing further work either on the part of Sun/SDC or NSA had been highlighted. Our evaluators commended us on the job we had done and privately shared their amazement at how much work they realized it had been for us. For me, for the LTD, it was done.

There is more work to be done to win NSA: proposal clarification, minor changes to bid items, &c. Some handwaving will be appropriate as well.

But we did the LTD in a way that we all can be proud of. We did it starting with a foundation of solid Sun products. A team was built, mostly comprised of Sun technical support engineers who weren’t afraid to roll their sleeves up to get the job done. Add to that some exotic hardware and software products and a manager - me - who I hope made a difference along the way.

Goodnight, that is, as soon as I unplug the alarm clock…

–Roger Frank

A few days later, Phil Wright, Sun’s Director of Federal Sales, wrote a letter to me, copying Sun management, thanking me for the HPW effort. It seemed a fitting conclusion to what we all had done. Here is what he wrote:

TO: Roger Frank
FROM: Phil Wright, Director, Federal Sales
DATE: 25 September 1986

CC: S. McNealy, B. Lacroute, J. Roebuck, R. Bik, L. Hambly

There are three vital steps to winning the NSA business; Preparing an outstanding technical proposal, Conducting a “Flash” and Compliance demo, and Pricing to win.

I would like to recognize the achievements of Roger Frank in leading the charge which has positioned Sun to win a program which has the potential for doubling Sun’s revenues and dealing a serious blow to Apollo.

I remember standing before the E-Staff and describing both the task, which was overwhelming, and the type of person needed - a mature, disciplined, focused person who could plan, implement, and com- mand the respect of Sun’s top technical resources.

One name was set forth, that of Roger Frank.

Few people are given the opportunity to play a pivotal role in changing the future for a company such as Sun and fewer still succeed, but I can say without equivocation that Roger Frank is such a man.

He and his team created a flash demo of such quality and substance that we intend to use it as the key demo in developing future DOD business.

Having achieved a major success in the flash demo and without respite Roger and his team undertook the compliance demo which required the creation of over 300 individual demos of functions proposed in our NSA bid. The execution of these demos took two twelve-hour days and we met the most rigorous requirements I have seen during my career.

Roger left his family and friends, gave two months of his life and worked 16-20 hour days to move us one very important step closer to a victory at NSA.

Roger left as he came, without ceremony, so quietly that myself and one other formed the thought in our minds and on our lips in unison, “Who was that masked man?”

Thanks, Roger!

Sun management recognized the HPW bid for what it was: not just an initial $287 million order but a deal with the government worth billions in sales and, more than that, likely a winner-take-all position in the workstation market. They were right. Sun won the HPW in September of 1986. By the end of 1987, Sun was first in the workstation market and Apollo was third.3 By 1990, Sun was running at $2 billion/year and Apollo was well on its way to becoming a footnote in computer history.4


  1. Before my career at Sun was over, I would attain several additional clearances to allow me to be the interface between Sun and certain other high- level government projects. During this part of my career even my family didn’t know what I was working on or even, sometimes, where I was.

  2. I must have been pretty close. In the hallway after the Flash demo had been delivered, an evaluator quietly commented to me: “How did you know?”

  3. In order of market share, largest first: Sun Microsystems, Digital Equiment Corporation, Apollo Computer, Hewlett-Packard, IBM.

  4. Between 1985 and 1989 Sun was the fastest-growing company in the United States, according to Forbes magazine, with a compound annual growth rate of 145 percent.