16 April 2006

Free in many ways

Pick a site with 1000 active computers in it — admittedly, that tally is to keep the numbers simple and understandable, but the principle works in the same pattern for 20 or 50,000 or however many machines — and say that you replace one machine with a Linux box, set up with a little thought. One of the workhorse machines from Motium — with dual disks — is quite a reasonable choice, and say you then disappear for three years.

The other machines, typically running MakeBux4BillO$, will require repair of some sort (e.g. active spyware or virus cleaning, finding lost docs etc) every few months on average. Say that JRandomTech charges AUD$80.00 an hour and a visit typically takes him 2.5 hours to survive the experience (call it 3 billed hours or AUD$240.00 per visit on average), that’s AUD$920.00 per annum per machine. 1000 machines adds up to about AUD$920,000 per annum.

Say that your machine runs for 3 years without manual intervention (with automatic updates, I have had several servers do over 5 unattended years securely and reliably, then email me a warning before any issue becomes a risk to them), it is now worth AUD$920.00x3, or AUD$27,060.00 simply for having not died (and remember to factor in wages etc not lost due to downtime, and more not effectively lost to diagnosis or arranging repair-people). The dual [RAIDed] hard disks help to make this practical — and with 3”-style drives, fast and painless to replace.

Standard, interchangeable hardware run by reliable software is simpler and easier to manage, wall-to-wall, than special-case “glue-ware”. Sports-car-embedded managers may not notice this at first, but the bean-counters certainly will. The numbers grow too fast to be safely ignored.

“You” can buy, build and install one of these for well under a grand, but consider the jaw-dropping prices of what each will be replacing and you might wind up with a good markup and massive up-front price advantage (for the customer) together. Another thing to whizz past your grey matter is the possible replacing of a great hulking (noisy, power-eating) metallic monster or rack-of-same with one or a few quiet, efficient, reliable replacements. Motium do reliable machines roughly the size of a sandwich and it’s generally not hard to make a Linux server 10x as efficient to run as the machine it’s an upgrade for.

If you replace 20 machines, then beancounters or the managers relying upon them can purchase cars or other rewardishes worth about AUD$81,000 more than the old ones, for a nett gross delta of $ZERO to the company. Attractive? That’s cost for nearly two complete brand-new and shockingly performant Peugeot 407, for example — a car that will blow the doors off almost anything worth up to 1.5x as much, and off many worth 2x or 3x as much, not to mention steering, handling and stopping unbelievably better than generally expected. Or get a Ford Fairlane 190 and a bit if immediate oohs and ahhs of recognition are more important than awesome practicalities and sustainable, ultra-safe unbeatability.

Oddly enough, good hospitals brought this all into focus for me, because of what I did not need to do while healers worked to keep me on-line (initally, “on-line at all”). This, self-evidently, is not a trouble-free universe [understatement] — but far, far fewer things required seeing to (and so far trusted others have dealt more easily with those things than with ten times as many old-fashioned “standard” [oft-preferred word choices: “unimaginitive, risky, unreliable”] engineering alternatives would have been susceptible to).

So if anyone asks you what FOSS has to do with hospitals, the instant answer is “Make them more useful, efficient, safe and effective” — and if you have trouble with the term “safe” you (or your audience) truly need to spend a little time witnessing what our medical staff really and competently field and work with in essence: all day, every single day. Even the merely “simple” technical issues can be mind-blowingly complex, then these people also mark added and serious social and administrative limits competently.

Doing this all through busted or dodgy computers is, to put it very mildly, not at all helpful.

These people each need a good heart and spirit, and it’s quietly stunning to me how many actual and excellent people truly qualify for the post and also for many surprisingly closely related posts. This technique is a simple way to minorly but directly help them and (carefully done) assure and relax their bosses.


Leon Brooks said...

By the way... smacking one’s head into a footpath at well over 10m a second (even via a stack-hat) doesn’t do it an outstanding amount of good. Please check my numbers until about August instead of just accepting them.

Kristy A. Bennett said...

Oddly enough, good hospitals brought this all into focus for me, because of what I did not need to do while healers worked to keep me on-line (initally, “on-line at all”).

on-line or saw-tooth-line? as opposed to flat-line...

Love this post, I think I may add it to my 'Reasons for FOSS Portfolio'.

Leon Brooks said...

Thanks, Kristi!

Kristy A. Bennett said...

FYI Have linked to this post on my site.


Leon Brooks said...

That's "on-line at all."

I think Joondalup Hospital were (for them) very quick to ship me off to Royal Perth because they didn't want my corpse on their books.

Good choice, too. Before we got to the freeway (a whole 3 blocks off) the tech had noticed that I wasn't breathing, so had fitted a tube.

BTW, on examining my email, I still get notified by at least 3 different antique machines that all is well.

What's special about this is that I haven't clapped eyes (or hands, or SSH clients) on them for over seven years.

I know that two of them would still be updating themselves if updates for that variant of a Linux distribution were still available.

I also know (have done similar) that I could clock on with SSH & get the things completely updated (12 whole distro versions, in at least one case) in a couple of hours.

Having a live Windows server survive 7 years on-line & unattended would involve radiant lights & winged beings (-: The hardware fails first :-)

Am I joking, there?

I've repaired three of my Linux servers so far that have had the CPU fans melt down completely after some years.

The servers still ran just fine. They were a tad hot to the touch, however.

In another case, a Linux server complains to me that it's not able to back up another (Windows 2000) server. Why?

Because the other server was off-lined over 2 years ago. Meanwhile, the site still gets email through and serves names, files & web pages from a physically identical (same model number) Linux server.

It's Free software, all right. Failure free! (-: