Microsoft Corp. has moved its focus away from sponsoring studies that compare the total cost of ownership for Windows and Linux, and is now turning its attention to reliability, an area where the perception favors Linux.
The Redmond, Wash., software company on Wednesday will release a reliability survey it commissioned and paid for, which finds that Windows Server 2003 is more reliable and robust and allows IT administrators to execute various tasks more quickly than those using Red Hat Inc.s Red Advanced Server 3.0 running on the same hardware.
The full study is expected to be available here on Wednesday. Microsoft has also posted its own interview with Katrina Teague, vice president of solutions for VeriTest, which performed the study.
Martin Taylor, Microsofts general manager of platform strategy, told eWEEK in an interview that there is a pervasive perception that Linux is far more reliable than Windows, and he wanted to see if that was in fact the case.
Touting his commitment to “getting the facts,” Taylor said he asked VeriTest, a division of Lionbridge Technologies Inc., to undertake a “completely independent study of the issue” for Microsoft.
Taylor said Microsoft paid for the research because “if we didnt, it wouldnt get done. But that does not mean the report is not independent or that the results are not valid.”
VeriTest measured the time it took a group of IT administrators—18 Linux and 18 Windows who had passed a screening process—to execute various tasks associated with improving the reliability and robustness of back-end infrastructure and end-user services in Windows and Linux production environments within a simulated medium-sized business.
VeriTest configured test environments with three Hewlett-Packard Co. ProLiant DL380 G3 servers running as an infrastructure server, e-mail server and file/print server. One set of test environments ran Windows Server 2003 and Exchange Server 2003, while the other ran Red Hat Enterprise Linux AS 3.0.
“The test environments were also specifically configured in a failure prone state. The systems were functional, but lacked basic hardware/software fault tolerance, up-to-date patches, and data access security,” the test report states.
The administrators were given a series of proactive and reactive tasks and spent 26 hours over four days on these. The proactive tasks ranged from configuring new devices and printers to implementing system backups, system monitoring and remote access.
As the administrators executed these tasks, a VeriTest test proctor initiated reactive events like device or system service failures that simulated typical system problems and required troubleshooting to resolve, the report said.
VeriTest captured timing and task completion results from a variety of sources including administrators journal files, instant messaging logs and system service probing script log files, and exit interviews.
Next Page: Initiating downtime.
Page 2
During the test, VeriTest also initiated a series of events that broke or disabled various system services in the administrators test environments, which remained down until they were fixed by the administrators. “We characterized this downtime as service loss and used a set of service probing scripts to measure the amount of end-user service loss time caused by the events,” the report said.
Each service loss event targeted a specific system service like e-mail and the printer, and caused that service to become unavailable to the user population. The probing scripts recorded when the service loss was initiated and when it was subsequently fixed by the administrator.
“Over 90 percent of the IT administrators felt the test environment was realistic and accurately reflected a real world IT environment when asked in exit interviews. The test found that the Windows Server 2003 environment had 4:20:19 of average end-user service loss time compared to 4:59:44 of average service loss time for the Red Hat Enterprise Linux AS 3.0 environment on measured service loss events. Lower results are better,” the test report said.
Under the test scenario, 32 more tasks and events were completed in less time in the Windows Server 2003 environment—280 events in 18:43:22—than in the Red Hat Enterprise Linux AS 3.0 environment—248 events in 27:47:46.
But the one major area where Windows Server 2003 administrators spent significantly more time reacting than did the Red Hat administrators was on mail server events. However, the study explains this by saying that only two of the Windows IT administrators said they had experience with Exchange Server 2003.
Microsoft, however, is not just turning to sponsored research in its quest to dampen the ever-growing interest in Linux. It is also sponsoring targeted open-source conferences and has made some of its source code available to some customers, partners and its Most Valued Professionals.
Last year the company also launched its “Get the Facts” campaign, which is designed to give customers information about the advantages of using its Windows operating system over Linux, its open-source competitor. But many of the statements and “facts” have been challenged by the Linux and open-source community.
Microsoft also last year launched a new multimillion-dollar advertising program for its Windows Server System, which it said would put a more human face on the campaign and the product.
In addition, it announced in 2003 a global initiative to provide governments around the world with access to Windows source code under its Government Security Program, designed to “address the unique security requirements of governments and international organizations throughout the world.”