Peter Houston, Senior Director, Windows Server Product Management Group
REDMOND, Wash., Dec. 17, 2002 — Businesses are demanding more reliable methods of quantifying the cost and overall value of adopting Linux products compared to Microsoft offerings. In response, Microsoft commissioned IDC, a recognized leader in cost analysis methodologies, to conduct a comprehensive study that evaluated the total cost of ownership (TCO) for Microsoft Windows 2000 and Linux server offerings.
The study evaluated Microsoft and Linux in the context of five specific workloads common to IT infrastructures today. IDC's findings suggest that Windows 2000 offers a lower total cost of ownership than a Linux-based server across four of these five workloads.
The results were recently presented in an IDC white paper, titled "Windows 2000 Versus Linux in Enterprise Computing: An Assessment of Business Value in Selected Workloads," which analyzes data collected from mainstream customers with hands-on experience using both operating systems. PressPass spoke with Peter Houston , senior director in Microsoft's Windows server product management group, about the significance of the study results.
PressPass: What motivated Microsoft to commission the study by IDC?
Houston : Customers have been asking Microsoft to help them answer an important question: Does the "free" price of Linux translate into lower long-term cost of ownership? This isn't a simple question to answer, because a lot of the components of TCO are not obvious. To help answer this question, we commissioned IDC to conduct a comprehensive, objective, third-party study around TCO for Windows and Linux.
PressPass: How did IDC approach the study, and why did it focus on TCO specifically?
Houston : TCO encompasses both fixed and ongoing expenses over a given period of time. It is one of the primary methods that businesses use to weigh potential purchasing decisions, yet we had not yet seen a rigorous TCO methodology applied to a comparison between Linux and Windows. We asked IDC to specifically evaluate areas in which Linux is generally perceived to be strongest in terms of total cost and adoption.
IDC compared five common enterprise computing workloads deployed in the Windows 2000 and Linux operating system environments: file serving, print serving, networking infrastructure, security applications and Web serving. IDC used a five-year timeframe, which is a typical lifecycle for many IT implementations of server operating systems, and interviewed IT executives and managers at 104 North American companies in multiple industries who have used both operating systems.
PressPass: What were the primary results of the TCO study?
Houston : IDC's findings suggest that Windows 2000 offers a lower total cost of ownership than a Linux solution across four of the five workloads most common to corporate IT environments. In the areas of network infrastructure, print serving, file serving and security, participants reported between 11 to 22 percent savings over a five-year period -- primarily as a result of Windows' significantly lower IT-staffing costs.
For Microsoft, the findings show that our investment in resources to engineer products that are comprehensive, easy-to-use, and deliver value "out of the box" ' for key IT scenarios is paying off for customers in the form of lower TCO. Interestingly, even some representatives from the open-source community have said they aren't surprised by the study's results.
I think that IDC's findings also underscore the fact that software acquisition costs account for only about 5 percent of TCO. This indicates that factors such as application availability, IT-staff-skill requirements, company standards, and long-term supportability have far greater impact on the overall cost of IT solution ownership.
The only area in which Linux offered a minimal cost advantage over Windows was the simple Web-serving workload. I'd like to add, however, that Internet Information Server Version 6.0 (IIS 6.0), which is included in the upcoming release of Windows .NET Server 2003, should close this gap due to significant improvements in manageability and hardware utilization.
PressPass: If Microsoft commissioned this report, how valid are the results for the IT community?
Houston: Microsoft asked IDC to run a careful, detailed and objective analysis of TCO comparisons between Linux and Windows 2000 across common IT workloads. The emphasis was on IDC to deliver independent, critical results for both operating systems. We can't speak for IDC; however, they have been very up front about the thorough, platform-neutral approach and overall research methodology they used to achieve their end results. We're confident that this study is an objective look, with tangible results reflective of real IT conditions today, and we're eager for others to take a critical look and prove the results for themselves.
PressPass: What is the significance of the findings for the enterprise market?
Houston: The findings of this study help to counter the perception that because Linux involves minimal up-front expense, it is less costly overall. This means that customers don't need to make tradeoffs between cost and business value, as I believe they do when they consider Linux. In short, with Windows, customers can obtain both greater business value and lower cost compared to Linux. We hope to see enterprises take a closer look at the business value they require from their IT solutions and Microsoft's commitment to delivering this value to an enterprise platform -- above and beyond our cost advantages.
PressPass: What do you think are the key differentiators between Windows and Linux indicated by this study?
Houston : This study underscores a couple of points. First, because it delivers value right out of the box, Windows enables administrators to be more productive. This means customers need fewer resources to operate and maintain a given number of Windows servers compared to the same number of Linux servers. Second, the study shows that, in several workloads, customers typically purchase additional products for the operating system, which drive up the software costs of a Linux alternative beyond what they would pay for Windows Server. So, in that sense, the Linux operating system component may be "free," but the total cost of software required for a given scenario can be significant.
PressPass: Why did Linux outperform Windows 2000 Server in the Web-serving workload area?
Houston : It's important to note that Windows only trailed in this area by a minimal cost difference. Additionally, the overall TCO of both the Linux and Windows alternatives were low enough that the cost of the software and the hardware (that tends to be slightly higher for Windows) were "amplified" and outweighed the people-cost advantages of Windows. As I mentioned earlier, we expect that IIS 6.0 will have lower hardware requirements and simpler administration than IIS 5.0 (which was measured in this study) and the gap will close.
PressPass: How does TCO fit within the array of factors that IT professionals should consider when making decisions about server system implementations?
Houston: While TCO is only one of the factors that should be considered by IT professionals when making an enterprise system decision, it is a high-profile and significant one. In a time when most companies are strongly focusing on their bottom lines, overall cost is even more important to them. Evaluating TCO gives an organization a point of reference for determining whether its IT activities are running as efficiently as possible.
Another important factor that is very difficult to quantify is business value. Customers are telling us they want solutions with built-in value that allows for heterogeneity and that makes IT easy to use, deploy, implement, maintain and grow to meet evolving business needs. These needs will become even greater with the development and deployment of Web services. The broad range of Windows offerings require less effort, customization and maintenance for customers, which results in the greatest possible business value at the lowest long-term cost.
PressPass: How have Microsoft customers and the industry as a whole responded to the IDC study?
Houston: I would say that there has been fairly broad consensus in the analyst community that customers shouldn't justify Linux decisions based solely on acquisition cost. Since the supposedly lower cost of Linux has been the number-one reason given by customers for their interest in Linux, the IDC study is helping broaden the discussion in terms of what really constitutes long-term business value for IT systems.