First Start giving away Ref. implementation free
Did not work
Then give away free with Solaris
Did not work
What did not work? Turn the entire company around instantaneously?
Releasing the app server supposed to take BEA and IBM and wipe them off the face of the earth? Releasing OpenSolaris supposed to make all the shops on the planet dump their current investments and infrastructure and load OpenSolaris?
Does it even occur to you that what they may well be planning is something long term? Something farther reaching than a 1-2 year plan?
Why release the app server for free? Because it encourages development on J2EE. It lets someone to google "J2EE", see "sun.com", click on it and voila! There's an app server to develop on, and even deploy on. That makes it easy for folks to try J2EE. When folks try it, some might even like J2EE. If they like J2EE they might even like SJAS.
It is in Suns direct interest to promote and accelerate adoption of J2EE in the market place, whether they inevitably get the direct business in the end (i.e. selling a SJAS license/support contract) or not. If someone goes to Suns site, downloads SJAS, writes an app, and then deploys it on BEA, Sun still wins.
It makes the platform more accepted, makes more people knowledgable on the platform, and expands the overall Java and JEE market. That's good for Sun.
Same with OpenSolaris. Sun basically had to choose to open it up, or abandon it completely for Linux. It's quite difficult to compete against free software for commodity services, particulary on commodity hardware. I think everyone recognizes how Linux/x86 was destroying Suns low end market. So, they've not only joined them by developing a kick ass commodity hardware, they've up'd the ante by freeing Solaris. Now the commercial market has a very viable alternative to Linux. FreeBSD was simply never in that hunt and has never had a large commercial entity backing it.
For those who want the percieved stability of a commercially developed and maintained OS, on commodity hardware, at a commodity prices (i.e. free or pay for service), Solaris offers that.
Now Sun can compete directly with Red Hat and IBM in that same space, whereas before it was shut out with pure SPARC/Solairs platform.
See, that's the key. Sun is on a level playing field in that space now. Before it was getting zero business, now it's getting some of the business. It doesn't need to be the dominant player, but it needs to be able to play. So, now they can play and fight apples to apples. You can't easily compare a low end Sun SPARC server to a Dell Intel server, but you can with their Opteron line. Apples to apples, bang for buck. Suns machines are leaders there now, even for customers that run Windows or Linux on them.
ThenGive away all software free and on top of it
No Services business
Have bad Tech support
Sell X86 machines
And then support 40,000 employees.
Does anyone see a business model here.
Sun has a huge Services business, just not the size of IBM. I have no experience with their tech support, but it doesn't sound any worse than any other huge organizations tech support. I'm rarely happy with anyones support.
Obviously if Sun has a miserable services business, and they plan on competing with a services business (which they are), then they're going to have to improve that services business in order to attract business.
First of all Sun is trying to make a failed commercial product succesful by making it open source. Lets assume it becomes succesful, will Sun continue to give it away free. Will if they do they cannot survive. The need to have some solid revenue source. I would rather support Apache product becasue I know it will always stay open. Or even support any product by IBM (I am not IBM employee or share holder) because I know they have money from services.
Sun has no choice but to leave it open. They CAN NOT take anything they've released under CDDL "back" from the community. They can withdraw from the community and support their own fork, but they can't pull back the actual software. Just like IBM can sell their own version of Apache, without making Apache any less "open".
Sun DOES have solid revenue sources from hardware, software, and services. They ARE profitable. Insanely profitable? Nope, but they're still turning the company around.
They are taking all of the doors that they can affect that prevent them from getting a crack at business and yanking it off their hinges.
People use JBoss because it's free and "good enough". People use Apache because it is free and "good enough". People PAY OTHERS to support JBoss and Apache. Is Suns webserver any better than Apaches? I don't know. Most in the community don't know. Why? Because it is has been difficult to evaluate, much less use, hard to compete with Free. Well, that's no longer an excuse. Now it can be evaluated on more even basis with Apache. Apple to Apples.
By releasing all of their Software, the software can be marketed to CIOs and IT managers from both directions -- the top by Suns sales force, and from the bottom by those who just may download the software, try it out and, perhaps, even like it.
Again, expanding the markets in which Sun can compete. More market, more opportunity. They certainly need to monetize that market, and that's the next phase, but with new hardware and wider exposure to their software, they are well positioned to leverage these new opportunities.
But it's not going to happen overnight. They're still laying all of the ground work. Since pretty much their entire stack is free now, it's almost all in place. Next is Niagra and the reduced power initiatives, then save for opaque internal changes that we can't see, they'll have everything in place and able to leverage and take advantage of it.
Will it work? Dunno. Nobody has ever done this before. We'll just have to wait and see.