|The IBM/Sun rumour mill
||[Mar. 24th, 2009|04:05 pm]
As a Sun employee, I cannot comment on the viability of the proposed acquisition of Sun by IBM. My position of an employee does not entitle me to access any information before the investing public do, so I'm in the dark as much as the rest. So, I will not be commenting on that aspect of the rumoured buy-out.
However, as a general thought experiment, I'd like to explore some of the hypothetical ramifications of such a move.
You see, all sorts of bloggers like to follow different trains of thought - one that has been forwarded to me is the sentiment that the move would let "the Linux community will instantly suck every valuable line of code out of Solaris, leaving only the rubbish behind". For some reason, this compels me to comment on all numbers of levels.
Right, so what happens if IBM absorbs Sun? We now have a company with a number of previously competing products:
- Industry-level Operating Systems: AIX vs Solaris
- Various web server/application platforms
- Java implementations: Jikes vs Sun's Java</i>
In all cases, there are products in the same market - what you make of that fact is up to you, but my opinion is that if such a corporate glomping got the all-clear, IBM would want something for its money... and AIX does not have an opensource branch - you do not hear of unix enthusiasts getting wet at the thought of the next OpenAIX release (not that I know of many who would do the same for OpenSolaris, but at least there's the potential for them to exist).
In other words, the tricks, hacks and beautiful magic that happens inside of the operating system are accessible only to those in IBM.
So, the good bits of Solaris that we could (should) charge for are things like zfs (the bit that all the linux heads are drooling over), dtrace (excellent for debugging, take your view as high or as low as you want) and vnics (virtualised networking is a cool little piece of tech, especially in the virtualised world) are not going to be thrown away by IBM - they're probably going to take those features first and throw them into AIX to make it bigger, stronger, better...yada yada.
Of course, I'm blatantly ignoring the other market advantages that IBM would gain from that sort of move, like a unified java platform that they own, solidified dominance of WebSphere in the commercial, unix web platform market, and a 10% share of the server market.
But if it were purely for Solaris (which is doubtful), shareholders would be wondering why they'd suddenly throw the prime cuts to the open source community, and would lose confidence which would visibly affect share prices. IBM might have a good open source reputation, but they have to look after profits first. Yes, they have a strong business interest in linux, but AIX still gets them money, so they're not going to throw a massively-scalable OS to the masses without putting it in their own OS for a test drive first. The same argument applies to dtrace, for realtime analysis of pretty much any process (although MacOS X has it in there already) and probably to the network virtualisation project.
Basically, if Sun is acquired, things possibly get less open, because IBM grasped a concept that Sun had problems with under the current administration : get the money, then open source.