By Tim Landgrave
Over the last three years, Linux has emerged from a cult phenomenon driven by thousands of programming enthusiasts around the world into an enterprise-ready operating system. This has become possible because Linux’s destiny has also shifted from the dedication of those enthusiast programmers to a small number of commercial, for-profit companies that are focused on finding ways to build their offerings around the Linux system and the needs that Linux adoption uncovered.
But for Linux to continue its encroachment on the corporate data center, the new Linux stewards will have to deal with the issues that have surfaced as they try to create a lucrative commercial environment around it. Recent announcements from one of those Linux stalwarts have many companies wondering if the Linux movement can sustain its move toward commercialization.
Red Hat is red-faced looking for green backs
Red Hat, the largest distributor of Linux, recently made some significant changes in its licensing and support policies, which have forced many large Linux customers to reconsider their Linux options. According to Netcraft, in just the top three hosting companies (rackshack.net, Rackspace, and Interland), Red Hat servers host over 250,000 sites and 460,000 host names. This doesn’t count other hosting companies, corporations, and other Linux professionals that host their sites using Red Hat versions of Linux. Unfortunately, the market share that Red Hat enjoys hasn’t led to significant profitability. The cost of maintaining and supporting both its base distributions and the enhancements that have made Red Hat the dominant Linux distributor also make the red ink flow freely on Red Hat’s income statement. If Red Hat is to survive as an independent Linux distributor, it has to generate sufficient revenues to be profitable. Without additional licensing revenues, it has no other place to turn. It has limited consulting capacity to generate revenue and certainly can’t compete with other companies that provide Linux consulting services, like IBM Global Services.
Red Hat’s announcement adds considerable cost to both its core distribution and to the annual support fees. These key hosting companies are able to provide hosted Red Hat Linux servers for around $1,000 per year, for which they generate roughly $200 to $300 in gross margin. This includes about $50 per server in Red Hat licensing and support fees. With Red Hat’s new licensing plan, that $50 becomes a $350 support bill, effectively making Linux hosting on the Red Hat distribution unprofitable. Certainly, Red Hat will do some volume discounting to allow the largest hosting companies to generate higher margins, but it still stands to lose one-third to two-thirds of its profits based on these licensing changes alone.
What's a Linux hosting company to do?
Linux advocates will say that Red Hat’s move is insignificant because hosting companies can simply shift to other distributions. While this is technically true, companies will still have to pay for the difference in the quality of support they receive from other companies vs. what they receive from Red Hat. Either way, their costs go up to some degree. Moreover, potential price increases from Linux hosting companies to cover the additional licensing and support costs will make the costs of hosting and supporting Linux closer to the costs of using Microsoft’s Windows Server 2003 operating system. Hosting companies that have favored Linux for the "free ride" that accompanied it are in for a huge pricing shock over the next 12 months while other companies seek to create profitable ventures from what has been a predominantly "free" operating system.
Why Red Hat matters
Many Linux advocates who are appalled by this "money grab" by Red Hat have been very vocal about their new distaste for Red Hat. Some even go so far as to suggest that Red Hat has outlived its usefulness. But they fail to understand the importance of a healthy company like Red Hat for the entire Linux industry. Without strong, independent Linux distributors, the hardware companies that have taken over Linux stewardship have no reason not to begin creating versions of Linux that take advantage of their specific hardware features and configurations. Releasing this code back into the public domain as required by the GPL has little effect on them because users would have to be able to duplicate the hardware features to take advantage of their enhancements. In the end, the fate of Linux begins to look a lot like that of its older brother UNIX. The resulting fragmentation would leave companies with no choice but to purchase all of their hardware and software from a single source—a specific hardware provider.
The company that is most likely to win in this scenario is IBM. With IBM and Red Hat the most likely losers if SCO succeeds in its current lawsuit, IBM could choose to "own" Linux now in one of two ways. First, it could buy SCO outright and exert its claim on Red Hat, owning both the Linux IP and its largest distributor. Second, it can wait until Red Hat goes bankrupt either because its pricing drives customers away or the cost of defending against the SCO lawsuit drives it out of business. Either way, hosting companies, corporations, and IBM competitors have to begin evaluating their options while the largest independent Linux distributor—and in fact the entire Linux distribution model—finds out whether there’s a place in the value chain as Linux usage continues to grow.