This month, Sun Microsystems announced a new entry-level server running Linux and based on an Intel Pentium processor. At the same time, IBM announced a Solaris-to-Linux migration program that seems designed to steal Sun’s customers.
In response to customer demand, Sun has announced that, rather than selling its own SPARC processor, it will sell low-end servers based on Intel’s Pentium chip. It will also run open-source Linux not just Solaris, its own operating system. The impact of these decisions is crucial to Sun’s survival as an independent company in the long-term. If Sun continues down this path, it lays itself open to becoming just a commodity vendor that will have to work hard to differentiate itself from others – Dell will undoubtedly be able to match or beat it on sheer price comparisons.
IBM’s aggressive Solaris-to-Linux program is a challenge to Sun, as it aims to dislodge the Solaris grip on certain segments of the market. In a platform war, where all participants are offering the same operating system, IBM clearly feels that it can win.
Sun is in a difficult position. If it doesn’t offer Linux when its clients want it, then it risks defections to its competitors – hastened by moves like IBM’s. But if it does, then it risks losing its differentiators and being forced to slash margins and compete on grounds that it will have difficulty winning.
In the past, Sun has been the ‘sports car’ vendor of the IT world – selling to a tech-savvy audience who are looking for a technically excellent product, perhaps requiring a little more skill to use than the standard offering – but worth it. By contrast, IBM and HP were the luxury saloon purveyors.
Ovum sees a future where Sun will offer a top-end ‘sports car’ solution for its technically demanding audience, and commodity servers for the rest. Its top-end solution could be based on Solaris and SPARC – or even Solaris and Intel’s high-end Itanium chip. It would appeal to companies such as financial services, CAD/CAM, aerospace and telcos.
Sun will want the ‘halo effect’ of its top-end solutions to increase the attractiveness of the lower end – at least to loyal customers at the top-end. If this doesn’t work, Sun could return to how it was in the mid-90s, before the rise of the Internet, e-commerce and Java made it the Web server de rigeur. If it does work, Sun will maintain its volume but look rather like Compaq did just before HP acquired it. Not a particularly desirable outcome. It seems that Sun’s main hope is that the HP-Compaq merger is so protracted and difficult that HP loses ground to more focused and internally cohesive competitors.
It’s not all bad. For users, Sun’s announcement is a further vote for Linux. It is likely that, within months, more new Linux packages will be shipped than all other Unix systems added together. On top of this is all the old PCs that have been converted to ‘cheap-and-cheerful’ Linux servers because they won’t run the latest version of Windows. The industry will benefit from a strong alternative to Windows and Linux is the only non-Windows game in town.