% fortune -ae paul murphy

Why open oource works for Solaris

Last week talkback contributor Anton Philidor re-iterated his argument that open sourcing Solaris reduced Sun's revenues without providing significant offsetting value. Here's part of what he said:

One major impact on licensing...

... is the reduction in revenues ... Developer loyalty will follow on widespread adoption by potential customers. No matter how interesting the operating system, will developers devote the years necessary to create products which have little market?

There are several kinds of answer to both halves of this.

On the most superficial level, it's not obvious that open sourcing Solaris has reduced Sun's real or potential revenues. There are two reasons for this, first they didn't make much selling it to begin with: my $49.95 Solaris 2.7 CDs got installed on a half dozen production systems ranging from an Ultra2 workstation to a v880 - and so did the Solaris 10 set. Secondly, the numbers are just starting to firm up, but right now it looks like roughly one support contract gets generated for every fourteen free downloads -and that's usually business Sun would not have had without the open source effort.

A more satisfying class of answers is based on the "upward averaging" we usually see when proprietory technology is newly open-sourced. Basically what happens is that good ideas migrate in both directions: improving both the formerly proprietory product and the earlier open source products. Thus proprietary Sun breakthrough products and ideas like those in Dtrace and ZFS migrate to BSD and Linux while small systems ideas like skinny variants and easy installers migrate to the Solaris community: broadening and deepening both communities - and ultimately giving Sun both better products to sell and more people knowledgeable enough to buy them.

The best set of answers, however, combine speculation about how the industry will evolve over the next few years with a carefully calculated strategic response. From sun's perspective the ten year technology threat is coming from IBM's takeover of Linux coupled with the explosive power of the cell grid technology. Thus getting close to Microsoft on inter-operability while giving Linux developers and afficiondos an easy migration route to a more advanced, and less constrained, Unix gives developers an opportunity to hedge their bets while leveling at least this part of the playing field for 2007 and later years.

The biggest, most obvious, and generally in your face, strategic enabler here is, of course, Sun's community development license: a have your cake and eat it too deal for people wanting to sell proprietory extensions to open source code. What's behind that, however, is a hidden follow-through: anything written for Solaris 10 is easily reproducible for sale to Linux and BSD users. Unix, after all, is Unix no matter who writes the kernel- meaning that IBM will face software and services competition from Sun developers just as eager as anyone else to get their hands on Cell based PCs and Servers.

Paul Murphy wrote and published The Unix Guide to Defenestration. Murphy is a 25-year veteran of the I.T. consulting industry, specializing in Unix and Unix-related management issues.