After years of struggle and fading recognition, it seems the end is finally here for Solaris and Sparc, the Unix operating system and RISC processor designed and championed by Sun Microsystems and inherited by Oracle in 2010.
In a move that will win it no PR points, Oracle sent out recorded telephone messages to employees who were let go on the Friday before the Labor Day weekend. Yes, firing by voice mail. Classy.
The exact number is being debated, but talk on one message board puts it at 2,500 positions cut. That’s both Solaris and Sparc engineers. The workers affected are primarily in the Santa Clara area, in a former Sun office not even a mile from Intel’s headquarters, but they are in three other states and India as well.
Then on Monday, former Sun executive Bryan Cantrill wrote a blog post called “The Sudden Death and Eternal Life of Solaris” in which he wrote that the cuts are “so deep as to be fatal: the core Solaris engineering organization lost on the order of 90 percent of its people, including essentially all management.”
This isn’t exactly a shocking development. Back in January, Oracle laid off 1,800 workers, a tiny number relative to its size, but it included 450 workers from the company’s hardware group and reportedly half of the Solaris division. Layoffs have a habit of coming in waves, and last Friday’s was the second wave.
Also in January, Oracle changed the release road map for Solaris. Instead of Solaris 12, it switched to “Solaris 11.next,” a rolling release that would be pretty much security fixes but no new features or advances in the OS. The same thing happened with the Sparc line, with Sparc next replacing planned chip upgrades and featuring less ambitious improvements to the line.
Sparc and Solaris won’t disappear overnight, and Oracle has promised to support both until 2034. But the two will likely be long gone by then. The bulk of the server hardware Oracle sells are x86-based running Linux. In a commodity world, proprietary hardware and software has little future.
Oracle gave Sparc and Solaris a good try
But this, as Cantrill said, looks fatal. And after seven years, I can’t say I blame Larry Ellison and Mark Hurd. They stuck with Sparc and Solaris longer than anyone else would have. Had IBM, Sun’s original suitor, taken over the company in 2009 like it wanted, Sparc and Solaris wouldn’t have lived to see 2010.
The fact is Ellison and Hurd were patient, very patient, but in the era of the cloud and reduced spending on data centers, not to mention a free, open-source Unix derivative vs. a highly custom one and the advent of x86 as a truly viable mission critical architecture, Sparc/Solaris didn’t have a prayer.
And all told, Oracle really didn’t get a lot out of this merger. Simon Phipps, former Sun executive and now a managing director of Meshed Insights, recently posted a long list of all the things Oracle has fumbled since the purchase. They include:
- Java was described as the “crown jewels,” but the real reason for buying Java SE — trying to sue $8 billion from Google — has failed twice.
- Ellison said Java’s role in middleware was the key to success, but Java EE is now headed to a Foundation.
- Oracle criticized Sun for “failing to monetize” Java (ignoring the fact Java made the market that Sun monetized with hardware in 1996-2000) and proposed a freemium model that’s not resulted in revenue.
- They embraced NetBeans, which is now donated to Apache (a good thing).
- Bureaucracy over MySQL security fixes led to a decent portion of the user community going over to Monty’s MariaDB fork, enough to start a company around it.
- Ellison said he would rebuild Sun’s hardware business and has made numerous updates to the Exadata and other lines, but its long-time boss John Fowler quit last month.
- Oracle decided to cancel Sun Cloud and dismantled the ready-for-cloud features of Solaris, then the market went cloud.
- Oracle completely bungled StarOffice, and its user base jumped to LibreOffice.
Some of this couldn’t have been foreseen, like the impact of the rise of the cloud. But other things were in Oracle’s hands, such as the fumbling of MySQL and StarOffice.
I’ll give Oracle a lot of credit for trying, not gutting the company, sticking with the Sun IP for a long time and really getting Java back on track after it turned into an unholy mess of security exploits. But in the end, we judge companies on results, not efforts, and the results aren’t that good.