The study, conducted by Giga Research and paid for by Microsoft, compared the likely costs over four years for a dozen medium-size and large businesses that were developing Web-based portals. Giga examined the costs of creating a portal using Microsoft's Windows operating system and related development tools as compared with those of Linux-based systems using Java 2 Enterprise Edition (J2EE) tools.
"The primary conclusion of the study is that Microsoft offers a substantial cost advantage over J2EE/Linux as a development platform for the applications considered," Giga analysts wrote in the report, which is set for release on Tuesday.
Microsoft plans to use the study's findings in its effort to convince customers that Linux is not necessarily a cheaper alternative to Windows, once all costs have been taken into account.
"I want to make sure companies have facts in front of them," Martin Taylor, the company's recently appointed general manager of platform strategy, said. "Every time I see a study that says customers think Linux has a total cost-of-ownership advantage, that's a sign people aren't seeing this information."
However, critics of the Microsoft-backed study noted the small sample size in the Giga study, which polled just 12 companies, seven of which were running Windows and five of which were running Linux.
Giga analyst John Rymer said surveying just a few businesses made sense, given that the research firm was trying to get an in-depth look at the costs involved.
"For this kind of a study, what you want is a real deep look at a relatively small number of companies," Rymer said. "Frankly, it's just hard to do 50 of those (interviews) in any reasonable time frame."
Rymer noted that the Giga study findings don't necessarily indicate that Microsoft's software is cheaper than Linux and Java in all cases, just for the portal-creation task it studied. Rymer said it is conceivable that Linux and J2EE could be more cost-effective for other types of applications.
IBM said the study's findings contradict what its customers are seeing, and noted that Linux continues to gain market share rapidly.
"The growth in Linux is amazing," IBM spokesman Ron Favali said.
Favali said server figures released by IDC last week and new data from research firm Gartner show that sales of servers with Linux are growing faster than with Windows. Last quarter, shipments of Intel servers running Linux rose 58.5 percent from a year earlier, compared with a 29.9 percent increase for Intel servers on the whole, according to Gartner figures provided by IBM.
Separately, Microsoft's Taylor took issue with recent pitch by Big Blue, in which it urged customers to consider consolidating Windows servers onto IBM mainframes running Linux. Contrary to IBM's claims of better performance, Taylor said Windows-based servers outperformed the IBM mainframes by 25 percent to 300 percent in various file-serving and Web-serving benchmark tests conducted by Microsoft.
"(Mainframe) performance isn't what people claimed it would be," Taylor said.
Talkback
Here we go again....
Typically, a rule of thumb would state that measurements taken in any well defined environment would yield well defined results.
In the real world, where each and every programmer has to rely on their environments, the freedom that Linux and the GPL offers is unquestionable. The capacity to utilise, without any restrictions the efforts of individual contributors to an open process does no less than cut "unnecessary efforts"/replication down to a bare minimum. And this is across the board. Why re-write communication logic -(eg VoIP)- for an app, when there are many OpenH323 applications to learn and tap from. Why allow yourself to be chained down to mere APIs when you can strip and redesign -if you so choose- aspects of Clustering algorithms to fit your purpose-OpenMosix springs to mind here. Why waste the time taken to administer an association with a proprietary distributor, with license restrictions and added admin costs, when you can start work immediately. Why wait to be audited?
In addition to the above, with free access to all levels of operation, Open source creates a better knowledge base than any restrictive operation could hope for. This better knowledge base does/will in turn produce/"continue to produce" a more dynamic environment within which to work.
Face it, those who work with open source are a great deal more intelligent about OS operations and their working environments than Microsoft Workers at similar levels, and they are so without having to give themselves big titles derived from Multiple choice questions that are targetted at pre-pubescents. One thinks here of horses with blinkers.
A fundamental fact that will catch up with all people witnessing this transition, is that, in real terms, proprietary releases, much as those released by MSoft, are boring by comparison to work with - much like a rat in a maze or one on a wheel.
The true test of an OS would be this: "If you are strapped in to the seat of a craft about to be lanched in to space, and you glance around and see a placard on a pannel that says - ALL SOFTWARE FOR THIS MISSION IS PRODUCED BY MICROSOFT (trademark)-in case of problems contact your Microsoft Cetified Engineer onboard" - what would you do?
When one sits to compare Microsoft products to the scripting environments of the *Nix, one can only think that maybe the analysis was done on apps that are similar to "Hello World". Could we have a list of the products that were compared please???
I've seen cost cases both ways Linux is cheaper than Windows and vice versa. Anyone know of other, objective cost stuides?
As usual, the study doesn't report the cost- bias, most don't.
Their studies include the costs for training and time involved in learning Linux as a native Windows user. Increasing numbers have been learning Linux/Unix in the classroom and this training and time expense does not apply.
What many studies fail to report is the rising number of people who started their training in the Linux/Unix environment. Many countries are using Linux in the classroom as the first O.S. the children learn to use.
[What follows is personal commentary--probably off topic]
********************************************************
This is not necessarily because they want to avoid "the evil empire" but because they are trying to build an knowledge-based economy. People who learn Linux/Unix first are going to find programming and software developement more intuitive. For these countries, it is the "fast-track" into a far more lucrative economy than one based on natural resources (like mining, lumber, agriculture, or even manufacture).
We are starting to see results now, but the real results will be seen within the next seven years or so.
Of course, as a person from the U.S., I see this as a looming finacial disaster--far greater than the loss of cheap oil or any other resource.
It will result in the eventual export of the entire U.S. economy, not just jobs. The U.S. economy is knowledge-based, yet many teachers are technology illiterate, as well as history illiterate, and illiterate in English--their supposed native language. One percent or less of the population is actually producing the wealth for the nation. Most all of the other endevors are effectively subsidized--it is believed that biotechnology will take the place of the computer oriented knowledge workers.
The irony is, of course, that this leaves even greater numbers of the population behind. While it is possible to retrain for computer science--I do not believe it is possible to re-educate an adult to be competent in the bio-sciences.
There is also the resistance to these sciences from an emotional context--the perception of the western world as "the great Satan" will be much more virulent with a focus on biotechnology. As far as holy wars go, "we ain't seen nothin' yet".