Linux's total cost of ownership (TCO) is a hot topic which thaws more than a few penguin supporters. People who are passionate about one technology or another are always ready for a fight.
Remember all the aggro about the truth of vendor-sponsored research last year when Microsoft commissioned IDC to evaluate the TCO of Microsoft Windows 2000 and Linux server offerings? IDC found that Windows 2000 offered a lower TCO than a Linux-based server across four out of five workload scenarios.
Pragmatic IT managers will look past all the paranoid and apocalyptic crap that goes with the Linux v Unix v Windows debate and think about cost and performance.
Press releases give one view of the truth, and I have just received a nice one from Oracle titled “Cairns City Council saves up to 75 per cent in software, hardware and maintenance costs running Oracle on Linux on Dell”.
I guess these tantalising savings estimates, expected over the next two or three years, are calculated in comparison to some “proprietary” alternatives. Another view of the truth is offered by true believers such as James Turner who writes for LinuxWorld in the US and reckons that there’s not much of a future for “commercial dialects of the Unix operating system”. He argues that hardware vendors’ ability to tune and customise their kernel in order to optimise their machines will not be enough of an advantage long term. He also claims that buyers can still have support headaches even with a one-stop OS/hardware shopping experience, and that the argument that Linux does not bring a commercial support channel and someone to sue when things go wrong is shot down by the likes of Red Hat and SuSe’s enterprise offerings. Is it? See Lee Anthony’s letter on opposite page. By the way, Mark Hall on page 21 believes that open source databases will be the death of Oracle and DB2.
And then there’s truth to be gleaned from good old reader surveys. In a survey of 291 IT managers US Computerworld found that 77 per cent of respondents were “extremely” or “very” reliant on Unix. Some 56 per cent said Unix would indefinitely own the high end, and another 24 per cent saw its importance declining but not disappearing. In another two dozen interviews, the surveyed IT dudes said that while they “loved the economics of Linux on Intel at the low end, they’re acutely aware that it’s still years away from the power, scalability, stability and support their data centres require”. Obviously, these results do not necessarily translate here with local enterprise IT usually more “low-end” in scale than it is in the US.
The truth is evolving and currently as clear as mud.