Before the days of planetisation, a small group of marketing professionals and engineers at Sun Microsystems Inc made up what was fondly called SunOS development. It was the well-known and accepted Berkeley System Distribution, which had become the environment of choice for technical workstations. This lasted for many years, until the big bang sent Sun into many new directions, flying through the galaxies. In 1987 the first big bang exploded and changed Sun’s operating-environment strategy. Sun would join forces with AT&T Corp to create a single Unix standard based on AT&T’s Unix System V release 4, System V.4. Changing the Sun operating environment was a major technological evolution, with high expectations of return, upon completion. As could be expected, the endeavour to switch to a System V base from Berkeley System Distribution was much more difficult and took longer than originally anticipated. Finally, after two and a half years and at least a year late, Sun released its first developers’ version in November of 1990. The product, code-named Zeus, was a basic System V.4 without the bells and whistles of SunOS. As could be expected it was only mildly successful for those daring few who wanted Sun’s System V.4. Of course it was a disappointment, but it was not surprising that Zeus fell short of Sun’s quality standards. Why did this happen? Market pressures, compounded by all the promises in 1987 forced Sun to release something, which was almost nothing.

Smaller planets

Zeus was a hold-over until a quality product could be built, but no one could accurately predict when that would happen. Over the next couple of months, the second big bang hit Sun. From what was heard, during the holidays, the Gods of Sun, namely Scott McNealy and his chief officers all got together, locked themselves in a room and conjured up the new Planetary System. The plan was to create a group of subsidiaries they would call planets. Each planet would revolve around the Sun at its hub, like any planetary system. The two most important planets would be SunSoft, which was not named at that point and which would handle the migration to System V.4, and Sun Microsystems Computer Corp, which would maintain the hardware business. Many smaller planets, not as strategic, were created to manage other software such as compilers, networking and integration tools. There was a clear mission for Sun Microsystems Computer to maintain the constant growth of hardware that supported the company, and it would continue to be directed by McNealy himself. SunSoft, although not yet easily focused, would be headed by Ed Zander, vice-president of marketing for the ‘old’ Sun Microsystems, a man who had successfully marketed Sun to the leading position in the workstation industry, it was generally believed. If anyone could do it, it would be him. When Sun announced the formation of the planets, the entire company went into reorganisation shock. This was far beyond anything anyone expected, including me. But, believe it or not, it came together very quickly and it was beautifully orchestrated. Each week exploded with news of who had received ‘what position’ in ‘what planet’. By this time it was clear that everyone mightily desired to be in the software division, SunSoft. It finally came time for my level, while everyone was politicking for SunSoft. The excitement around the company was electric.

In the beginning there was Fire, Earth and Sun Microsystems Inc, a single entity with a vast amount of resources. Somewhere along the way the Gods in their heavens decided that planets were needed if Sun was to prosper in the highly competitive computer universe. Our special correspondant Bentley Radcliff, once of SunSoft and presently working at Apple Computer Inc on the Macintosh emulator, took out his stardate logbook for us and looked back at the emergence of the planetary system.

Opportunities brightly flashed in front of us. Those of us with courage jumped on them, landing firmly in the SunSoft camp. Now all that needed to be done was to figure out what our jobs were. Th

ose first few weeks in SunSoft were full of golden opportunities, complexities, new issues and political nightmares. The first job to tackle was a ‘perfect name’ for the company. Everyone became involved. Zander had asked us to submit names and said he would choose from the best. I remember the exciting meetings with the creative director, Zach Nelson, a highly motivated person, who really understood the Sun system. He would conduct brainstorming sessions with many of us to identify names. During one of the sessions Zander burst in, dropped a few hints of what he liked, he had obviously already made up his mind, and left as abruptly as he came. When we finally got around to a trademark name search, nothing was legally available. Anyone who ever participated in a name search will know that is common. I sometimes feel that there is not a single name in the world that is not already used, much less the great ones. Zander’s favourites were unavailable. But with courage and a staff of attorney’s anything is possible. So Zander got his way. The original plan was to name the company Solaris, which means Sun in Italian. Another choice was to call the company SunSoft. It would finally come down to making a decision and that was up to Zander. He liked them both so he decided that we would use SunSoft as the company name and Solaris as the name of the product. His gutsy comment, I remember to this day, Let ’em take us to court, but I am getting those names. The names were established and it was now time to work on the announcement. Someone came up with the idea to re-announce SunOS as Solaris 1, even though it had been out for years. We would then lay out a plan for the upgrade to System V.4 as Solaris 2 – that was actually quite brilliant. By doing it this way SunSoft would build brand acceptance for Solaris long ahead of final release date. By this time we were ready for a mission, which was at the top of our wish list. The obvious approach was to ensure Sun’s migration to System V.4, but the company needed more. Sun had in recent years spurred a small but growing clone market and had provided SunOS for compatibility. All the years that Sun had shared its technology with the world, it held back a few critical keys to make itself appear better than the clones.

Jockeying

But the mainstream idea behind the creation of SunSoft was to give ‘Sun Clones’ a chance to steal a march on the opposition. This was critically important if Sparc was to stay a leading architecture in the market. So the decision was to build a mission based on the migration to System V.4, and to create an equal playing field for all clones. Reasonable, and quite a big mouthful, but Zander had more up his sleeve than just Sparc. Early on he had decided that to be a volume player he would need to expand his newly named Solaris to other systems. Most importantly iAPX-86 from Intel Corp. And that is what he did. While all the planning was going on behind the scenes, the program development rolled ahead. The engineering organisations had to break apart their old structures and reform into a new unit. The jockeying of people always has an effect on performance, but those people really pulled it together. The biggest problem management faced was the years of engineering loyalty to the hardware division. Could SunSoft really provide a equal playing field? Time proved that the answer was ‘yes’ but it was tough climbing to that plateau. By the time all the fussing was sorted out, Solaris was well publicised. The focus was to ramp up Solaris 2 development, while maintaining the Solaris 1, that is SunOS, quality. This meant that new releases had to be built simultaneously. The environments were vastly different but had to run in parallel, which was actually quite a feat. Plus the engineering teams had to develop the tools to assist in a clean migration. In the marketing division, product plans and requirements were being redrafted and submitted to the engineering groups almost daily. We formed strategic committees to review and decide what was important and what got trash

ed. In three months of combined efforts with engineering we created a software development framework that helped us finalise the functionality and features in the first Solaris 2 release.