Supporters of Linux cried foul
this week when Microsoft
rolled out research reports that it commissioned from research firms IDC, Giga Research and the META Group that questioned total cost of ownership (TCO) of the open source operating system.
In the process, analysts involved with the reports are finding themselves under fire over them, including one more than a year old, because they say Windows enterprise server environments are less expensive to maintain than comparable Linux setups.
The reports are the linchpin of a new marketing effort by Microsoft, called “Get
the Facts on Windows and Linux” that contend that Linux isn’t as cost-effective just because it is essentially free software.
A Microsoft spokesperson said over the past year, software cost and value have been a common issue raised by IT customers. “Our customers have told us they want research and information to help make value-based IT decisions. The ‘Get the Facts’ advertising campaign aims to bring some of this information to companies who are making decisions about their IT solutions.”
The ads on some technology sites are designed to drive customers to the URL (http://www.microsoft.com/getthefacts) where they can find third-party evidence on what customers and analysts are saying about Microsoft Windows vs. Linux and total cost of ownership, the spokesperson said.
TCO is a business indicator that assesses the price tag of a product, from
the staffing costs to training to downtime. For software vendor marketing
teams, it’s still more popular to point at return on investment (ROI), or
the amount of time it takes to make money on the initial investment in the
product, as an indicator of a product’s worth.
But the reaction in the open source community was deja vu for Al Gillen, IDC systems software research director. As one of the authors of the Microsoft-sponsored report, “Windows
2000 Versus Linux in Enterprise Computing,” he said he fielded a lot of
phone calls in November, 2002, when the report was first released.
Despite the second round of contention over his report’s findings, he told
internetnews.com he is sticking by the results, which he said were
meaningful when the information was collected in 2002. Gillen said it’s to be
expected that a report conducted more recently would have different results, but that doesn’t invalidate the previous findings.
“Pull out the pricing for the enterprise-grade products from (commercial
Linux distributor) Red Hat and put it next to Microsoft and do a comparison,
you’re not finding a 10- or 100-to-one ratio at this point,” he said.
The price of Microsoft licensing is pretty insignificant in the total cost
of an enterprise switch, or even an upgrade, Gillen said, so comparing it to
the free or less-expensive alternative of Linux is only telling part of the
He pointed to the contested report, which said the cost of software is only a fraction of the TCO; staffing and hardware eat up much more of an
enterprise’s budget. Then there’s the matter of training an entire IT
department to become proficient in the Linux OS, which takes time and money.
“There’s always care and feeding that goes into supporting an operating
system over a long term,” Gillen said.
Commercial Linux vendors certainly know that. Both Red Hat and
Germany-based SuSE, in the midst of a Novell acquisition, charge for
technical support for their Linux distributions. Still, SuSE General Manager for the Americas, Holger Duroff, told internetnews.com that doesn’t make Linux deployments more expensive than Microsoft.
“I don’t agree with the findings at all. There is training to bring new
customers into Windows, as well as when Microsoft customers upgrade,” he
said. “If you compare the cost for upgrading a Linux solution to Windows,
Linux comes out better.”
Officials at the Open Source Development Labs (OSDL), home of Linux kernel
creator Linus Torvalds, couldn’t be happier to see all the interest being
drummed up by the Microsoft campaign. Nelson Pratt, OSDL marketing
director, said the anti-Linux campaign by Microsoft, as well as the recent
ad campaigns by IBM, only give Linux more credibility in the market place.
Pratt doesn’t think the results by firms like IDC and META Group are in
error, but that reports of any kind are skewed.
“Studies are what studies are, and they look at things with a very specific
eye with a very specific measurement criteria and you end up with statistics
that support whatever angle you are looking to support,” he said.
“Interpretations of what is the total cost of anything is based on a certain
set of criteria that is deemed to be worthy of measure. As soon as you
extrapolate beyond that study, the criteria are weighted and all kinds of
calculations can be made.”
Pratt said Linux isn’t geared towards competing against Windows in the first
place; most companies looking to adopt Linux come from the Unix world, an
operating system which holds the lion share of the enterprise server market,
upon which Linux was first derived.
“Linux’ heritage is actually much more aligned with Unix than Windows and
we’re finding that a lot of the end users we deal with in our User Advisory
group are really talking about Linux as an alternative or extension or
addition to do Unix deployments,” he told internetnews.com. “The
choice is not Linux versus Windows, it’s Linux as an addition to the Unix
Laura Didio, an analyst at research firm Yankee Group, said
although Linux works well for smaller
companies with a tech-savvy IT department, the cost of installing the
operating system in a large enterprise setting is still prohibitively
She’s putting the finishing touches on her own Windows/Linux TCO report, due
out next month (and not sponsored by anyone, she emphasized), which will
conclude as much. If anything, she expects Linux pricing in the enterprise
environment to increase in the next five years, when one factors in the
increased costs of interoperability, warranties, integration and technical
services and support.
“I don’t think Linux is going to end up any cheaper than Windows, in fact,
in certain environments it’s going to be more expensive,” Didio said. “In
other words, users are going to find out that there’s no such thing as a