Saturday Nov 24, 2007

A Fundamentally Flawed Strategy

Here are some interesting comments (and my reactions) from a piece in Network Computing Magazine -- Will Sun Shine Again? -- where Sun's John Fowler, EVP of Systems, was interviewed on a variety of issues. I just picked out the OpenSolaris bits here:

IBM has embraced Linux and Microsoft OSes as well as it's own OS stable all of which has paid off for the company.

Sun's response to this desire for openness wasn't to embrace Linux, but instead to create OpenSolaris. Fowler says the goal wasn't necessarily to get others contributing code to the Solaris kernel but instead to create a conversation about the source code.

Sure, we didn't necessarily need code contributions to the kernel because the code base was already mature and stable and being developed by 1,000 engineers around the world. However, we always wanted code contributions, and we always wanted those contributions to represent new ways to use and extend the system. In other words, we wanted to grow in new ways that didn't necessarily represent Sun's core markets. And we started that process by opening the code and engaging in conversations about the code. So, Fowler is absolutely correct. And I agree, too.

He also says that it'll take a decade for strategy to prove itself, so we shouldn't judge it yet. Be that as it may, the notion seems fundamentally flawed.

Fundamentally flawed? Actually, the strategy has been quite successful even in these early years. In just two years we've built a nascent development community, we are clearly growing globally, we are taking code contributions (and other contributions), we have an early governance model, we are opening our infrastructure, there are a few distributions based on the kernel, and now we are expanding the program even further with a new binary to engage not only new levels of developers but also users as well. And that's characterized as flawed?

If what you're looking for is a huge developer community that values the ability to see the source code for the operating system, Linux will obviously win over Solaris.

Why must one system live and the other die? Instead, why can't both thrive? Also, the "huge developer community" probably has as much to do with the binary as it has to do with with the source code. Some would argue more, actually. That's what Indiana and the other distributions are designed to address -- to engage users and application developers building on top of the system. The number of developers actually interacting at the kernel source level and helping to build the system itself is much smaller, and it will always remain much smaller. Regardless, our strategy always included a long term, phased approach of opening code, infrastructure, and people and building a multi-layered community around the core. This can't be done all at once. It takes time. What you see now is a snapshot in time. We were different a year ago, and we'll be different a year from now. So, again, I fail to see how this strategy is "fundamentally flawed." Also, I fail to see how any comparison to Linux (or IBM or HP, for that matter) makes any sense whatsoever. I think over time, people will come to realize that community building is not necessarily a zero sum game. There is room for diversity. The world is a big place.

And by opening the source code, Sun has both created a product in OpenSolaris that won't have a revenue stream and given its Solaris faithful a reason to look at AIX and HP-UX based systems.

Actually, the exact opposite is true. By opening the code, the Solaris faithful are sticking with Solaris, and some who left are coming back. Also, it was the "Solaris faithful" who wanted us to open the code in the first place. They told us this quite directly, in fact. But we didn't open the code exclusively for the faithful. We opened it to reach new people, too. And we are. By the tens of thousands, actually, and we've been able to do that because the code is open. And regarding revenue streams -- remember that OpenSolaris is a development project. Sun's product is Solaris. There's a difference. Sure, as the OpenSolaris binary distros evolve, I'm sure business models will emerge and Sun will participate as well. But for right now there's no reason to confuse the obvious -- OpenSolaris is a development project run by a community, and Solaris is a product supported by a company.

For some applications, that closed development process used by the likes of HP, IBM and formerly Sun, that results rock solid software married to rock solid hardware is desirable, if expensive. While there's no doubt that Solaris is still a solid OS, there is some doubt about where Sun sees its fortunes and its future for Solaris.

Just because we are opening our development processes doesn't mean we are throwing out our development processes. Over time, non-Sun community members will earn their way just as Sun community members. This is already occurring, actually. As far as the "doubt" bit, well, you can't convince everyone, I suppose. When I look at the massive engineering and business investment Sun is making in Solaris and OpenSolaris and new support from AMD, Intel, IBM, and Dell on top of HP claiming they sell more Solaris systems than anyone else, well, "doubt" is not the first word that pops into my mind. I can think of a few other words that come to mind, though. Can you?


« July 2016

No bookmarks in folder