Zum Erhalt Ihrer Wettbewerbsfähigkeit tendieren Unternehmen zunehmend dazu, die Kosten für den Betrieb der Server in Ihren Rechenzentren höher zu gewichten als das in der Vergangenheit der Fall war. Einer der Gründe für die Gewichtsverschiebung bei den Entscheidungskriterien sind signifikante Kostenvorteile durch alternative Plattformen. Wenn heute Betriebskosten vergleichen werden, beispielsweise von Windows Servern im Vergleich zu Unix-Servern, ergeben sich nicht nur marginale Vorteile – die das Risiko einer Migration keineswegs aufwiegen würden – sondern erhebliche Einsparungen.
Laut den Branchenanylsten von IDC führen die Ergebnisse dieser Neubewertung des IT-Betriebs zu spürbaren Veränderungen von Marktanteilen. In der Analyse “Understanding UNIX Migration: A Demand-Side View” vom Januar 2006 hatte IDC fstgestellt, daß 45% aller Unix-Migrationsprojekte mit Windows-Servern realisiert werden. Im Februar bestätigte IDC dann (IDC Presse-Mitteilung vom 22. Februar 2006: “Worldwide Server Market Slows in 4th Quarter but Grows to $51.3 Billion in 2005, Highest Revenue in 5 Years”), daß die weltweiten Umsätze mit Windows-Servern ertsmals höher waren als die Umsätze mit Unix-Servern.
Eine Studie (“Driving Lower TCO and Rapid ROI through UNIX Migrations“) von Mercer Management Consulting hat sich kürzlich mit den Hintergründen dieses Trends beschäftigt. Die Berater befragten im Auftrag von Microsoft 30 Entscheidungsträger auf Kundenseiteum herauszufinden warum Kunden von einer Unix-Plattform zu einer anderen Plattform wechseln, anhand welcher Kriterien sie Ihre zukünftige Plattform auswählen, und welche wirtschaftlichen Ergebnisse sie mit einer Migration erzielten.
John Wenstrup von Mercer Management Consulting hat in einem Interview mit Presspass weitere Einblicke zu der anhängenden Studie geliefert. Seine Einschätzung der momentanen Trends:
“In the mid-1990s, UNIX grew to become the dominant server operating system for business computing. It led in both market share and server revenue. The UNIX market still has a huge footprint today with more than 3.5 million servers installed and in use by customers worldwide. However, other platforms have been chipping away at that number pretty consistently. In fact, most estimates concur that as many as 40 percent of those UNIX installations will be moved to new platforms over the next two or three years. IT executives and industry pundits tend to converge on four main drivers of this transition:
- Cost vs. Performance. Competing platforms have made remarkable strides in performance against UNIX, such that many believe platforms such as Windows Server have essentially caught up with UNIX performance. Given the fact that UNIX is still much more expensive than competing platforms to purchase and manage over time, UNIX appears to offer a poor trade-off of cost versus performance for most application implementations in the marketplace today.
- UNIX End of Life/End of Support. Many UNIX vendors are ending support for older, legacy UNIX platforms, which has spurred organizations to consider a wider range of options in the marketplace. As one executive commented, “since the vendor essentially forced me to re-platform anyway, I took the opportunity to look broadly and chose Windows instead.”
- Intel-Based Server Success. UNIX vendors who have been pushing RISC-based UNIX servers for a long time have begun to support Intel platforms, particularly given the promise of robust 64-bit computing. As a result, IT executives have become even more convinced of the potential of competing platforms on Intel architecture.
- Improved Tools to Ease Migration Process. Conventional wisdom used to say that migrating away from UNIX would be time-consuming, complex, and costly. In fact, most companies report that migrations have gone far more smoothly over the past three years than in the past, due to better migration tools and processes. The result has been faster and cheaper migrations.
For these reasons, UNIX migration continues unabated and looks like it may even be accelerating in some areas.”
Zu den weiteren Ergebnissen der Studie meinte er:
“In addition to the finding that IT executives continue to seek opportunities to migrate servers from UNIX, we found that Windows is the preferred choice when IT organizations migrate servers as part of a focused effort to improve business processes, deploy critical applications or restructure the IT architecture – what we call “transformational migrations.” Another key finding is that Linux appears to have gained a significant portion of its traction in smaller, more tactical UNIX migration decisions, many of which are based on less financially rigorous analysis. We also discovered at least three items that we believe are common misperceptions about the UNIX environments and UNIX migration, as follows:
- Misconception #1: UNIX migration is hard. Our study showed very clearly that UNIX migration is not as difficult as many IT professionals expected. Robust tools, the knowledge and capabilities of today’s IT staff, and the way that many applications are architected make the migration much easier than many executives anticipate.
- Misconception #2: Platform costs drive the economics of a migration decision. When some companies are considering migration, they assume that the costs of the hardware and software platforms are the overriding cost elements. In fact, what we found is that most of the cost of migration – around 80-85 percent – is composed of the labor of executing the migration and the ongoing labor associated with managing the environment. The total cost of the entire hardware and software platform – including maintenance costs – tends to only be 15 to 20 percent of total migration costs. Even more interestingly, the upfront acquisition cost of the software platform – e.g. the operating system – is only about 2 percent of the total migration costs, and even with ongoing software maintenance cost, is only around 5 percent of total cost. Obviously, this has implications for a company like Microsoft that competes in a world where competitors such as Linux are perceived as providing lower costs.
- Misconception #3: Linux is obvious choice since it is “free or close to free.” A few individuals in our study said that they had chosen Linux as a no-brainer “because it’s free or close to free – or at least it’s cheaper for tools and the operating system.” Many others claimed that “it’s easier to migrate from UNIX to Linux because of the API (application program interface) compatibility between UNIX and Linux” or because “we can take advantage of our in-house UNIX knowledge in the Linux environment.” However, and quite surprisingly to us, the vast majority of companies who performed even moderately rigorous TCO (total cost of ownership) analysis made it clear that the risk-adjusted returns for migrations suggested Windows was often essentially a wash with Linux – and sometimes even advantaged – on total costs. As one executive told us, “we were, frankly, surprised that Linux and Windows provided about the same TCO, so we could make a strategic decision based on application availability and access to resources in the market.”
To summarize, migrations really are easier than most people think, the platform acquisition and operating system costs are a tiny portion of the overall migration cost, and deeper TCO analyses disprove the conventional wisdom that Linux is the “obvious” best, easiest and cheapest option.”
Viele Entscheidungsträger würden spontan vermuten, daß eine Migration zwischen zwei verschiedenen Unix-Derivaten (schließlich ist Linux ja letzlich ein Unix-Derivat) einfacher sein müsste, als ein Wechsel auf eine vollständige andere Plattform, wie beispielsweise Microsoft Windows. Demnach müsste das Vorhaben technisch riskanter, unwirtschaftlicher, und mit grösseren organisatorischen Anpassungsschwierigkeiten verbunden sein. Wieso wechseln dann aber trotzdem so viele Unternehmen – wenn man den Marktdaten von IDC glaubt – zu Microsoft Windows? Warum sollte beispielsweise ein SAP-Kunden beim anstehenden Upgrade auf Netweaver gleichzeitig seine Applikationsplattform wecheseln? Wenstrup beantwortet diese Fragen folgendermassen:
“Many of the companies considering transformational migrations, such as moving high-end workloads or strategic workloads for business applications, have found that when they conduct a risk-adjusted analysis, the Windows platform has a very attractive overall TCO compared with Linux. The study cites as typical the experience of a European telecom company undergoing a transformational migration for its ERP application. According to the company’s head of IT architecture, a risk-averse CIO wanted to lower IT costs by moving an SAP application from the back-end and integrating it into a front-end Web server. The CIO wanted to stick with UNIX, but the head of IT architecture convinced him to consider Windows. When they did an analysis, they realized that Windows was both the lowest total cost solution – even compared with Linux – and the lowest risk solution, and the migration to Windows was fully supported by SAP.
Of the 30 companies we talked to, many ended up moving to a Windows environment after doing rigorous analyses for risk-adjusted TCO, often to debunk conventional wisdom. Many companies not only did a TCO analysis and a performance analysis ahead of time but actually went back retrospectively and checked to see whether those numbers proved themselves. We did have a couple of companies that said they achieved all of the performance levels that they had anticipated, and they had either met or exceeded their cost numbers. Again, this has been borne out by companies in large-scale implementations today.
Our study participants also pointed out that the Windows platform is a lower-risk solution than Linux around a few dimensions. First, it offers access to skill sets and capabilities, both partner and internally hired. Second, it offers “future-proofing” around applications in particular, meaning it prevents companies from getting locked into a limited set of applications as their business changes. Lastly, the Windows platform offers robust tool availability for things like systems management and so forth.“
Im Interview betonte der Berater noch einige andere Aspekte die direkt mit den Folgen von Wirtschaftlichkeitsanalysen zusammenhänge, deshalb lohnt es sich, daß Ganze vollständig im Original zu lesen.