Linked by Thom Holwerda on Tue 3rd Apr 2007 23:18 UTC
Sun Solaris, OpenSolaris "Sun's CEO Jonathan Schwartz loves to splatter the media with the line that Windows, Red Hat Linux and Solaris stand as the only operating systems of significance in the server kingdom. We've spent the last few years struggling to appreciate the seriousness of that claim. Sun's declining system sales failed to inspire much optimism about the company conquering the data centers of tomorrow with a deflating 'venerable' OS. A couple of recent items, however, have tweaked our view of Schwartz's favored claim. It could well be that Solaris - of all things - provides the 'iPod moment' Sun seeks." In the meantime, Sun upped the speed of some of its SPARC chips.
Thread beginning with comment 227112
To read all comments associated with this story, please click here.
The turning point
by zizban on Tue 3rd Apr 2007 23:26 UTC
zizban
Member since:
2005-07-06

Two things really turned Solaris around: First was a full commitment to x86, instead of the lukewarm if-we-must attitude they had prior to Solaris 10 x86. They now treat x86 the same as the Sparc version and it shows.

Second was open sourcing Solaris; it wasn't GPL but it created a huge buzz and a flourishing community.

Reply Score: 5

RE: The turning point
by butters on Wed 4th Apr 2007 01:11 in reply to "The turning point"
butters Member since:
2005-07-08

Sun is the first major platform vendor to rediscover the fact that although investors and analysts mostly care about the whiz-bang hardware that drives the revenue for the vendor, developers and IT professionals care about the less glamorous operating systems that drive the revenue for the customer. Expect other major platform vendors that have traditionally been hardware-oriented to place more emphasis on their operating systems going forward.

What we have had in the recent past is hardware that was literally running away from the software--hardware that was getting too wide and too big for software that was designed to manage scarcity. Today the high-end is topping 128 hardware threads and 1TB of memory. The question isn't how to give each process its fair share of the resources, but how the heck we're going to keep those pipes busy and what in the world we're going to cache in that ocean of memory. How to we let software take advantage of these beasts, and how do we make them easy to manage?

As soon as CIOs began using words like utilization instead of bandwidth, the OS became arguably more important than the hardware. Mediocre OS vendors (i.e. Microsoft) began to lose market share (slowly but surely), and the hardware vendors that outsource their OS (Dell) became weaker. It's a whole new ballgame today, and the OS is on the pitcher's mound.

Solaris has a lot of things going for it, but Linux is going to be a challenging competitor unless Sun can expand the scope of Solaris to encompass more of the market. They need an attractive desktop OS and a powerful workstation OS to go along with the the server OS that sells the hardware and services. The OS is not a niche product that can be targeted to a particular market segment. It's a brand that is central to the vendor's systems strategy and a set of technologies that must be widely used and supported throughout the industry.

Consider the three major UNIX variants today. Linux runs on damn near anything, they give it away to whoever wants it, and it's equally suitable for everything from cell phones to mainframes and everything in between. Every qualified administrator and developer coming out of school today is familiar with Linux and its customary set of tools. Solaris runs on commodity PC hardware and a midrange server architecture, they have free and proprietary versions, and it's suitable for servers and workstations. The vast majority of admins and devs coming out of school are aware of Solaris and many have used it. AIX runs on midrange to high-end proprietary servers, it's licensed as a part of the hardware purchase and support contract, and it's suitable mainly for databases and high-performance computing. Many admins and devs coming out of school have never even heard of AIX and learn it on the job.

Linux followed the strategy of flood the low-end of the market where most of the mindshare is developed and hope the larger accounts follow, and they've been very successful. Solaris wants to appeal to the middle of the market where most of the server volume is positioned and hope the mindshare follows, and they're been pretty successful. AIX wants to lead the high-end where the profits are the biggest and hope the competition stays off their heels, and they hold onto a small lead in UNIX revenues.

What we have here are three distinct approaches, although Solaris keeps drifting closer to the Linux strategy as mindshare seems to be more important than revenues and as Linux is quickly gaining enterprise-friendly features such as virtualization. Now that HP-UX and the late Tru64 are falling by the wayside while SPARC is becoming weaker in the high-end and stronger in the low-midrange (Niagara), AIX and System P are basically competing with IBM's resurgent mainframe systems.

I think the "turning point" is not specific to Solaris but general to the entire IT industry. Before, there were various levels to the market that you could target individually. Today, you're either ubiquitous and driven by mindshare or a committed niche vendor driven by high margins. Sun found itself in no-man's land with Solaris, deciding to challenge Linux on its home turf (de facto standards) instead of challenging IBM's strength (technical leadership). They've got a long road ahead of them, but at least they stopped and asked for directions.

Edited 2007-04-04 01:16

Reply Parent Score: 5

RE[2]: The turning point
by kaiwai on Wed 4th Apr 2007 01:40 in reply to "RE: The turning point"
kaiwai Member since:
2005-07-06

Just a couple of points:

1) I agree with the push on the desktop, laptop and workstation; I've pushed an idea through the marketing mailing list for a "Solaris Workstation Edition" where by there are periodic re-spins every 4 months, and packages for the distribution are available on a respository so that upgrades and updates can be done without needing to download an entirely new iso. I also have pushed through the idea that ON builds, not only include sources but pre-build packages so that novice users, and developers who have little time to compile, can download the latest and greatest, and test it in every day user - thus expanding the pool of testers.

2) Desktop/Laptop/Workstation is at the heart of attracting developers; attract them with a good desktop operating system, and they will come, learn the operating system, become excited about the direction, and mindshare is added to the developer community.

There seems to be a disconnect that you're a developer and a user, and they occupy seperate spaces; what about the developer sitting there writing code who wants to listen to his mp3's whilst working? what about the developer who does some part time programming at home, but also likes watching DVD's and movies on his computer?

This is where Sun falls down, assuming you can neatly catagories people into pigeon holes, and they never leave them - all the developer does is write code; I'd love to meet a programmer who only doesn't programming on his computer.

3) AIX is going to hang around simply because there are alot of IBM shops still out there - generally speaking, if you're a big IBM customer, it makes no sense mixing and matching, you just go with them, and they'll provide you with a 'great deal'.

This will be further entrenched with the move to standardise their whole high end on a single processor, where by you'll have from mainframe to server all using the POWER processor - thus leaving it up to the customer to decide what is suitable for which job - but beyond that, I don't see AIX expanding much, it just sits there, adds little revenue to the bigger picture.

Reply Parent Score: 3