In October 2005, I discussed a Microsoft-backed study that compared the reliability of Windows Server and Linux ( See "Microsoft vs. Linux 2005 Edition: It's All About Reliability" at the second URL below.). According to that study, Linux systems were consistently more difficult to manage over time than Windows Server in a real-world e-commerce scenario. The study highlights a problem I've always found to be true with Linux: Although few should doubt that system's capabilities, the componentized nature of Linux makes it hard for any one vendor to deliver the same kind of integration and platform completeness that we take for granted in the Windows world.
Flash forward to today, March 2006, and Microsoft is facing off against Linux again, this time backed by an IDC study that demonstrates how Microsoft partners that earn a Microsoft Competency in Advanced Infrastructure Solutions or Information Worker Solutions tend to significantly outperform other Windows or Linux partners. And this is real world performance we're talking about: net profit margin, revenue growth, gross profit margin, and so on.
Ryan Gavin, the director of platform strategy at Microsoft, summarized the findings nicely: "Moving to Linux is not a panacea," he said. "There's no fundamental difference in profit margin for companies that primarily support Windows compared with those that primarily support Linux. However, the group of Advanced Infrastructure partners saw 46 percent higher profit margins than with Linux."
As Gavin noted, there's an elephant in the room as well. ISVs wishing to support Linux must make a decision. They can back a single Linux distribution--Red Hat, SuSE, Ubuntu, or whatever--and hope that version takes off in the market. Or, they take on the additional costs and complexities of supporting multiple Linux distributions and versions. The problem, of course, is that "Linux is not Linux," as Gavin says. ISVs have to target Red Hat Enterprise 3 or SuSE 9, or whatever. They aren't interchangeable.
This situation stands in sharp contrast to the opinion that one of Linux's biggest strengths is its componentization. But the reality is that no one company or organization truly controls Linux. Even a Linux-only company such as Red Hat is largely just a packager of others' work. And when it comes to delivering a solution on a platform as complex as Linux, companies often have to do the plumbing themselves, taking on additional cost and time themselves. What these companies really want to do is build solutions on top of a reliable platform that just works. As Gavin quipped, "they want to remodel the kitchen, not be the plumber."
Microsoft's approach to this is pragmatic. If you look at partner programs for Linux from companies such as HP, IBM, Novell, or Red Hat, you'll find that they're technology based, not business based. So, Microsoft even has a model advantage. On the Microsoft platform, you don't have to worry about first getting the IT stack to work.
And if you look closely at these Linux companies, some interesting facts emerge. Consider IBM's recent move to Linux, a decision that was based largely on the company's ability to generate income for IBM Global Services. By contrast, Microsoft isn't interested in making money from services, and the company instead uses its partners as its "services army," as Gavin called it. Novell, meanwhile, is a disaster in the making. Unlike other Linux companies, Novell had a ready-made audience of upgraders waiting in the wings, thanks to its Netware installed base. But those customers are mostly migrating to Windows and not Linux, because Windows is more of the traditional model that they're used to.
So what does this all mean? Customers choose Windows because it's a complete platform, and they can look out over several years on the roadmap and know how things will be improving. The purported strengths of Linux--its decentralized development model and componentized architecture--aren't translating into measurable benefits for those that back the platform. Suddenly, Windows is the safe choice. This latest study is just another step down a path that I think will lead to a logical conclusion. And for those companies that have standardized on Windows, it's all good news.
If you're interested in more information about this latest study, Microsoft has published an executive summary (in PDF format) on its Web site.
Microsoft Competencies: Partner Pathway to Business Performance
Microsoft vs. Linux 2005 Edition: It's All About Reliability