The sponsored research approach may not have played well for Microsoft, but IBM wasn't deterred this week as it trotted out two new sponsored reports heavy with praise for the total cost of ownership (TCO) benefits of Linux over Windows and Solaris.
The two reports, one of which was an update to a Robert Frances Group TCO report penned in 2002, were conducted by RFG and Hayward, Calif.-based Pund-IT Research.
Overall, the reports showed Linux had continued to lead in TCO benefits over the competition and that it had expanded into new roles in companies that had previously allocated Linux to so-called "edge" uses like e-mail and Web serving.
Pund-IT principal analyst Charles King, whose report focused on three companies using Novell SuSE Linux or Red Hat, found that Linux on the server had progressed into an arena of "second stage benefits." The three companies - pharmaceutical wholesaler Alliance UniChem CZ, Boscov's Department Stores in Reading, Penn., and Zahid Tractor and Heavy Machinery in Jeddah, Saudi Arabia -- each migarted to Linux between 2001 and 2002.
King described first stage benefits of Linux as those that were associated with the open source operating system today: lower cost hardware and minimal additional costs.
"The second stage benefits we found were a simple expansion of those benefits; that Linux provided a simple variation of IT efforts, enhanced IT staff performance and helped to retain existing IT staff," King said.
The one interesting point that struck King was that the three companies, which serve starkly different industries -- each adopted Linux for similar reasons.
"These are markets that are widely separated by geography, and despite that, all three adopted Linux for the same reasons -- to lower costs and increase overall IT efficiency," King said.
King said all three companies reported that they have reached or are exceeding the goals they hoped to attain with Linux, including better scaling of software and optimization across the value chain.
Linux goes long term
The second study, conducted by RFG's senior business analyst Chad Robinson, focused on larger enterprise-level companies and how the use of Linux had evolved since the original RFG TCO study in 2002.
"TCO and cost itself was a primary criterion for selecting Linux, but that has now changed," Robinson said.
Instead, enterprises are searching for the short-term and long-term value of Linux deployments, the transfer of skill sets from one system to another and the ability to obtain applications from multiple vendors.
Another change Robinson noted was how the competition is behaving, especially that company up in Redmond, Wash.
"Competition has increased sharply with Microsoft decreasing costs, like with its Web edition of Windows, and Sun has been putting out additional attention in the form of supporting Opteron and OpenSolaris," he said.
Robinson's research found that each of these competitive factors has narrowed the TCO gap from where it was in 2002, but customers still said Linux was significantly less expensive.
Another difference between the original study and this month's research was the rise in long-term Linux deployments. In the past three years, such deployments were seen by many users as experimental phases still out on the edge of the software stack.
"Now Linux is in a more critical applications layer, and our research looked at the applications server, focusing on J2EE [Java 2 Platform, Enterprise Edition]," Robinson said.
Another trend was a shift in administration skill set specialization, which had always been present in IT shops but has been increasingly looked at as a way to cut costs.
"Now there is an admin who is solely responsible for the OS, and nothing above it. There is a separate admin responsible for JDM and nothing in the outside world," Robinson said. "This consolidates costs, and allows for more productivity."
Previously with Linux this was not yet the case, because IT shops using Linux had either just finished experimenting with the operating system, or the Linux group was separate department altogether.
This specialized approach worked with Linux due to the flexible nature of the OS, Robinson said, as it allows an company to deploy it across x86, mid-frame or mainframe equally.
"The long-term roadmap for Linux is very dependable. It has changed from a few years ago when a significant question in industry was where Linux was going and how it was going to grow and compete," he said.
The specialized skill sets factor was also the one area in the report where Linux was not as successful in topping the competition. According to Robinson, roughly half of the companies contacted had still not implemented a specialized administrator skill set, but he added that trends lead him to expect that gap to close eventually.
To sponsor, or not to sponsor?
Scott Handy, IBM's vice president of worldwide Linux strategy, dismissed questions about the credibility of vendor-sponsored research, and said requests from IBM customers for research similar to RFG's TCO comparison have persisted since that time due to what he perceived as a more objective approach from IBM.
"I firmly believe customers find our approach more balanced due to the fact that we have built a healthy business around Windows, Linux and Unix," he said.
RFG CEO and chairman Cal Braunstein also chimed in with regards to sponsored research concerns that have arisen with the help of the often-criticized Microsoft "Get the Facts campaign." He said interest from users in the original 2002 TCO comparison between Linux and SPARC had remained constant up to the release of the August 2005 study.
This article originally appeared on SearchEnterpriseLinux.com.