Linux poised for move from data centre to desktop - report

Clearly viable on server, client to follow within two years

A report published this week says that Linux is now a credible alternative for the core of the data centre, and will be one at the client end within two years. These are perhaps not what you'd call revolutionary conclusions, but Butler Group's Linux in the Enterprise (further information here) is a fairly detailed and comprehensive survey of the state of art from a business perspective, and should be pretty useful reading for executives looking for background while they consider taking the plunge.

Hardware independence, the report reckons, is the most important attribute of Linux, giving customers more flexibility together with the opportunity to reduce the number of platforms supported. At the moment Linux's traction in the enterprise is most pronounced against Unix derivatives, but beyond this it could provide an exit from Windows (although Butler stresses that Windows remains the best strategic choice for many companies who're already using it). The hassles of a switch to Linux, the report says, are of a comparable level to a Unix version upgrade, or of a move within the Windows camp from NT 4 to XP, and Linux could prove particularly useful for the implementation of a consolidation strategy.

"An organisation that commits itself to just two OSs, i.e. Windows and Linux," it says, "will over time reduce the skill sets required for management and administration."

On the desktop Linux is viewed as viable to some extent, but open source office tools will be playing catch-up with Microsoft Office for the foreseeable future. Companies with a heavy investment in MS Office systems will also hit problems in terms of file formats and macros, and will find these difficult to escape from. But Linux is becoming a valid desktop alternative for those without a requirement for interchange with Microsoft Office.

And we particularly like its healthy scepticism on TCO studies: "Disregard the generalised industry studies that discuss Total Cost of Ownership (TCO) and whether Linux, UNIX, or Windows is most expensive to operate. TCO is complex with many variables to consider, making direct comparison difficult as each organisation is different. IT management must maintain its objectivity in selecting the most appropriate and cost-effective OS for the selected services and applications. As with any IT deployment, the choice hinges on each individual organisation's circumstances and what the specific business and technical requirements are."

Hoorah! Shout that louder, please.

Butler recommends Novell/SuSE and Red Hat as the distributions most worth considering for the data centre, but gives others a look-in on the desktop. Which is a conservative stance, but probably a sensible one if you're talking to people and companies without a great deal of open source knowledge and technical expertise.

And it's clear about the alternative to acquiring some: "For those organisations that can reconcile themselves with being locked into one vendor for the majority of the software stack then Microsooft is a valid option. What the emergence of Linux has done is to provide an alternative and while not yet up to the sophistication of Windows, it offers a flexible platform strategy that many will want to adopt over the next five years."

The report probably has far too much background and detail to be of direct value to those already committed to open source, but it's a useful primer for them to shove under the noses of the people they want to convince. ®

Related Stories:

European Commission publishes its guide to OSS migration