« Institute for the Future report on location based services   |   Main   |   Perils and travails of social networking »

Solaris 10 on x86 - critical to Sun’s continued success

The Register’s Ashlee Vance has a piece on Sun and its future - positing that Sun will only retain relevance if Solaris 10 on x86 becomes the new growth engine. Vance takes the audacious positon that Sun is in extreme danger of becoming “a bigger, richer version of SGI”. I definitely recommend you check out the full article, but it goes like this:

Sun has for many years garnered more than its fair share of attention and power for one reason and one reason only - Solaris. There are more Solaris customers on this planet than IBM, HP, Intel, Microsoft, and Dell care to remember. If HP came to you and said there are only three important operating systems on the planet - Windows, Linux and HP-UX - you would laugh. If IBM did the same with AIX, you would laugh again. When Sun pulls this trick with Solairs, the thesis flexes a bit of muscle.

This part is very insightful, because its at the crux of the fight between HP and Sun over Schwartz’s blog posts (which I blogged here) that marginalize HP-UX as a dead platform. But Solaris’ franchise is based on its success with the RISC architecture - which is itself in decline. Therefore, it is critical to Sun that they get their ‘best of breed’ operating system out there on as many x86 boxes as possible, as quickly as they can.

There are reasons to believe Sun can pull this off. Its Solaris engineering team is arguably the best operating system design group on the planet. Sun pours billions into the operating system and has countless features that neither Windows or Linux can match. It’s secure, scales incredibly well and now, Sun claims, fast. Beyond all of this, Sun faces no real Unix competition from a big player in the x86 market. It’s Sun’s segment to win.

Should Sun carve out a large chunk of the x86 market and have Solaris running on the systems, the rest of its business will likely thrive. Sun will sell more of its enterprise software - another growth area - and sell more storage. A thriving x86 business even legitimizes much of Sun’s Linux desktop play, which opens another avenue for growth.

Sun has a lot of interesting RISC technology on the way such as its multicore processors, high-end file systems and virtualization technology. This will keep its current customer base happy and might even help Sun regain some lost ground in the high-end market. No matter how well this business performs though it will not carry Sun to the $50bn mark.

Sun’s other bets such as making more money off of Java, thin clients and renting out computing power may pay off in the long run. They won’t, however, pay off like a massive Solaris x86 server business. That’s where the billions in upside are.

Interesting stuff.
Updates: So, while “El Reg” can bring a good perspective - they just don’t report on news like the other tech publications out there. Information Week has an informative rundown of what McNealy said, with more detail than offered by Ashlee:

Throughout the day, Sun execs sought to position Solaris as a low-cost operating system that’s a better choice than Red Hat Linux–the most popular version of Linux among corporate IT departments–for small computers. At the same time, Sun’s technical investment–the company spent more than four years developing the new system–would provide its platform a longer shelf life than Unix operating systems from IBM and HP, they said. Solaris 10 includes performance-increasing technology called DTrace that the company says is giving test customers 20% increases in performance over older versions of Solaris. It also includes a new file system called ZFS, security technology taken from Sun’s Trusted Solaris, which it sells to government accounts, and technology called Containers that could make servers easier to manage. Solaris 10 also includes the ability to run Linux applications with a small performance penalty, as a way to migrate apps from Linux to Solaris.

Sun also is offering its customers nontechnical incentives to buy its technology. McNealy said Sun will offer its customers indemnification against lawsuits arising from using its technology, a scenario that’s increasingly on the minds of corporate software buyers. “You, too, may have a Kodak moment,” McNealy said, referring to a $92 million payment Sun made to Eastman Kodak Co. last month to settle a patent infringement lawsuit over the Java programming language.

Sometimes the detail is nice too!