While some of the open source projects that Sun Microsystems created -- and which I used to work with -- have maintained a high profile, the one most associated with Sun in the minds of system administrators has been strangely forgotten. Whatever happened to OpenSolaris?
The OpenSolaris story illustrates a key value delivered by the liberties provided via open source. When your vendor changes direction, what do you do? Traditionally, you have five options:
- Reimplement your system to use your vendor's new strategy
- Reimplement your system using a different vendor's product
- Use your vendor's locked-in support contract
- Use an escrowed version of the source code (which you wisely demanded from your vendor during the original purchase) and engage specialists to carry on with your existing software at your own risk until it no longer works as the environment evolves
- Proceed apace and hope nothing goes wrong
[ InfoWorld's Simon Phipps pegs an unlikely enemy for open source: your procurement rules. | Track the latest trends in open source with InfoWorld's Open Sources blog and Technology: Open Source newsletter. ]
All of these are, to say the least, suboptimal. Most of the victims of vendor abandonment I've encountered have opted for one of the first two scenarios, some using either of the last two as a bridging strategy. All of them have found the experience very expensive.
While superficially appealing, the value of escrow as a safety measure is especially questionable today. It was a viable protection in the days where all the software on your computer was delivered on tape and compiled for your system. But now, with the best will in the world, the source code to a complex proprietary product won't be easy to engage even with plenty of warning and resources, let alone in an emergency on an unmodified budget. Finding anyone sufficiently familiar with the source code to navigate -- not to mention safely modify -- it is very unlikely, and doing so on demand verges on the impossible.
With open source software, there's a sixth option. As long as it's real open source and not compromised in some way, the community around the free-software commons can rehost and carry on. They can step aside from the original project host and follow the vision they collectively believe in. This can be a fork -- as it was for Jenkins and LibreOffice -- or it can be a straightforward continuation of a discarded initiative, as in the case of Apache River (Jini) or, notably, the heirs of OpenSolaris.
Technology lives on
To be clear, the OpenSolaris name is no more. As is obvious from such sources as staff blogs, Oracle has no commitment to open source for what was formerly the OpenSolaris platform, and it stopped using the name almost as soon as it took control of Sun's assets. While Oracle is willing to make available source code for a core subset of Solaris 11, it has no interest in delivering the full source or receiving anything more than bug reports from paying customers through formal channels. There's no role for community in Oracle's plans. Perhaps that's a reason customers are deserting the company?