Is Custom Linux the Future?

I have been reading quite a few articles about how China, among other countries, has been seriously considering (or actually pursuing) development of its own flavor of Linux, largely due to the expense associated with Microsoft server operating systems. While this revolution might not happen overnight, there are some interesting scenarios in which I can see large user communities (financial services, international governments, etc) banding together to develop industry or national versions of Linux.

Let me start off by saying that if this happens, it will not happen overnight. While Linux is definitely in strong growth mode, we are still talking about an operating system with (I believe) 5-15% market share in server operating systems (editorial note — there are several fascinating articles about the politics of counting servers vs. revenue and just what counts as a deployment). I think that there are two really interesting communities in which proprietary versions of Linux could really take hold in the not-so-distant future:

Computing Intensive Industries (Financial Services, Insurance, Telecommunications, and Life Sciences) – The incentive for this group is clear — these three industries spend quite a bit of money on Windows and big Sun servers. Interestingly, this group of industries has been one of the first to embrace Linux in a significant way. Also, financial service firms in particular already have large software development staffs who are doing custom application development — why not just repurpose some of those folks to develop Linux? I am certainly not the one to suggest that we will see the CIOs of Goldman Sachs, Morgan Stanley, and Merrill Lynch will sitting around the table hammering out the future of Linux. That group does, however, have a pretty significant financial incentive to move to commodity Intel hardware running Linux.

Foreign Governments – Countries such as Ireland, India, and China, which have become havens for offshore development for U.S. companies, have sufficient numbers of trained engineers and a local cost structure that would make such a development possible. One should not, however, confuse possible with feasible — as with any effort to develop a standard, there would no-doubt be significant bickering and wrangling as to what the “country standard” version of Linux should look like. One can only wonder, however, how long it would take several thousand engineers under the direction of a foreign government to develop a stable, approved version of Linux. All GNU license issues aside, I think this is a fascinating question.