Wrong Metric
|
| Author | Content |
|---|---|
| azerthoth Jan 17, 2007 5:24 PM EDT |
Is it just me or have journalists missed the boat on this one too? You cant compare apples and oranges, which is what your doing when your attaching a dollar sign to the linux market. Some starts using linux on a server and pays for the whole package, pre configured everything, bells and whistles, support, etc, etc. Then some bean counter wraps his fuzzy little brain around the fact that he can replicate the whole thing for nothing more than the cost of hardware. This is suddenly a 'hidden advantage' or 'competative secret' and gets abit of glossing over. Better yet, someone like our friend Helios walks in with distro in hand and just gets to work making the magic. So why the insane clinging to a flawed metric? |
| jimf Jan 17, 2007 5:30 PM EDT |
> So why the insane clinging to a flawed metric? In a capitalist world, Linux is the odd man out. |
| devnet Jan 18, 2007 5:36 AM EDT |
I think it's actually because businesses will be able to snow over everyone with Linux. There's a danger that Linux will get a bad name BECAUSE of businesses using "comparative analysis" and showcasing "hidden advantages" and so on...but they'll actually be snowing over their potential customers. It's sad because potential customers need to know the reasons Open Source is good for them and not the reasons Company X can provide a key to unlock the "hidden advantages" of Open Source. It's like getting a copy of the want ads and typing up a synopsis and offering it to people as a drop in replacement for want ads..."See? I've removed all the crap you don't need and have given you a competitive advantage over your counterpart corps!" But in actuality, they've just pulled the wool over the eyes of businesses who don't have the resources or time to effectively analyze Open Source software without first deploying it. |
| jezuch Jan 18, 2007 3:15 PM EDT |
"You can't measure it with dollars? But... but... but then it doesn't exist!!!!" |
| jimf Jan 18, 2007 3:27 PM EDT |
> "You can't measure it with dollars? But... but... but then it doesn't exist!!!!" Exactly! There are other reasons, but 'the main one' is that it's incomprehensable to them. |
| azerthoth Jan 18, 2007 3:40 PM EDT |
On another thought, just where were these folks that were interviewed? That could skew your results horribly in either direction. This has the flavor of having been done in North America/EU. Ask the same questions in South America and I'll bet your "poll" would have shown that Linux was already in place and that other OS's were the minority. So we have incorrect metrics based on a highly localized poll trying to pass itself of as a "global position". I should either stop reading or stop thinking, one or the other seems to frustrate me to no end. |
| pogson Jan 21, 2007 2:01 PM EDT |
Not the wrong metric, just not the whole of it. It would be a useful metric to calculate the value of Linux servers, but to do that one needs to consider that because they have longer uptime, Linux servers are more valuable than that other OS. (I remember the IBM ad of a few years back with suits walking through the computer room on tour only to notice things missing. The computer geek says they put everything in one server running Linux over there in the corner.) Server consolidation is not a sale, but one sale might suck up a lot of other hardware and processes. In my school we have Linux terminal servers serving 35 simultaneous users and they would top out at about 60 users. These things cost only $1200! Our web server does authentication, file service for the /home directory, firewalling, databasery and lots of other stuff and it is idling. These servers were all bought as parts and built here. We need to maintain them. Building them is a good start. We move many gigabytes every day for a few dollars. That other OS would cost us more for one server than all our servers cost and we would still need to buy thick clients instead of thin. Linux wins on the desktop as well as on the server here. Somewhere, somebody like Netcraft knows how many gigabytes/s are flowing to and from servers on the web. Usually it can be determined whether the server is Linux or that other OS. That would be an interesting part of the story. It would be interesting research to compare that with the depreciation of equipment and operating cost. In my place the ratio would be huge. My in-house stats (courtesy DarkStat): 112,718,050,299 bytes moved in and out of my central server. This box cost $1500 and I expect it to be lasting five years, about $300/year, or about $25/month. It uses 250 watts and electricity costs about 6 cents/KWH or $11/month. It cost about $500 to build it and to install the software, so the last month cost something like $8. So $11 + $8 + $25 comes to about $0.40 per gigabyte moved. The value of the services provided greatly exceeds that amount so the net cost is almost nothing. My maintenance costs are almost nil. Twice a year I will shake out the air filter. A few minutes a day I will spend marvelling at how stable it is and go on with my life. My server runs itself with scripting. With that other OS, I would have paid a licence fee for the server and each connected client, many thousands more. $1000 +$40 x 200 clients = $9000. The licences alone would cost more than $1 per gigabyte and worse, that other OS would likely have folded under the load. So the metric based on server sales is about a factor of three out even without considering the free nature of Linux (servers can be bought without OS). My terminal servers are marvels. I can have 1400 processes running on them all day long without a hiccough. I have seen that other OS choke with 100 processes running. I figure $50 worth of terminal server can easily handle one user. Over five years that is about $10 per year. Some organizations pay $1000 per year to maintain that other OS on a thick client. I generated 700 computer accounts via a script in five minutes and can install any software on three terminal servers in a minute by issuing a single command on the central server. Piece of cake. I think it is very realistic to mark up the value of one Linux server tenfold over that other OS. Sale price is a very small part of the cost or value of a server, especially with that other OS. |
| tuxchick Jan 21, 2007 3:49 PM EDT |
> Sale price is a very small part of the cost or value of a server, especially with that other OS. Great post. I get so tired of the bald-faced lying Microsoft apologists who say things like "Windows has better ROI and TCO because Linux admins get paid more, and migrations cost a lot, and Windows license costs are insignificant line items in an average IT budget." I just want to stab them between the head. A lot. |
Posting in this forum is limited to members of the group: [ForumMods, SITEADMINS, MEMBERS.]
Becoming a member of LXer is easy and free. Join Us!