The late August release by IDC of its quarterly server tracker data shows no let up in Microsoft (MSFT) Windows continued gains in the server market. Methodically it has overtaken legacy-operating-software-based servers and is now taking aim at UNIX/Linux- based servers, the category leader. According to IDCâ€™s reports, at 38.2% of 2Q07 factory revenue, Windows servers account for the single largest segment of spending, by operating system, in the worldwide server market. Windows gained 4.0 points of revenue market share over 2Q06. In my view (UNIX and Linux are one segment since Linux is UNIX), UNIX/Linux combined at over 40% still outsells Windows, but the comboâ€™s growth rate is much slower than Windowsâ€™.
By the IDC methodology, the Linux server revenue growth rate in Q207 inched back ahead of Windowsâ€™, a statistical metric in which Linux had led for almost 10 years in IDCâ€™s server tracking survey before Q107. In effect, the 10.4/18.7% and 10/19% quarterly growth rates (Q107 and Q207) for Windows and Linux respectively represent a statistical tie so far in 2007. Revenue for both types of servers is growing at three times the overall market average.
Microsoftâ€™s dominance of the server cannot yet be considered complete but the trend lines all run in Microsoftâ€™s favor. As long as Windows-based sales continue to grow as fast as Linux alone (which means much faster than the Linux/UNIX combination and which has been the case so far in 2007), Windows continues toward the 50% of sales mark formerly held by UNIX and MVS.
As I wrote in May, this also indicates to me that Linux as a server platform is leveling off much sooner than proponents anticipated. Linux-based servers are simply replacing UNIX-based servers, not Windows or legacy systems, in the normal multiple-year sales pattern. This also explains the push by Linux suppliers such as Canonical and Novell (NOVL) to get Linux onto desktop machines. Red Hat (RHT) rightfully seems to be focusing more on putting its whole stack, including its new development tools presumably, on the server. That move provides a powerful differentiator against other open source software (OSS) pure plays.
The leveling off of Linux growth as a server platform also makes OSS proponentsâ€™ emphasis on appliances and in Software as a Service (SaaS) environments more critical. The Linux on desktop effort is a lost cause on my opinion but the emphasis on appliances and SaaS, where the identity of the operating software becomes less important, is the next great computing market opportunity. Linux looks strong out of the gate.
Countering Linuxâ€™s early strength however in appliances and SaaS, as I researched over at ebizQ.net, is the fact that many new open source software (OSS) applications and other upper-stack software are being developed to deploy on Windows. Many others are being ported to Windows because of statistical phenomena such as the IDC server report. Itâ€™s not exactly a Catch 22 but it may have a similar effect.