Zonker's Blog

So, now that the dust is starting to settle from the news that Sun is buying MySQL for a staggering $1 billion pricetag, let’s take a look at the larger picture– how is consolidation going to affect the open source ecosystem? MySQL was the first project in 2008 to be absorbed this way, but it certainly won’t be the last.

The trend is clearly towards single-project open source plays being snapped up by larger players like Sun, Red Hat, Novell, and so forth. If you look at any company that’s backing a single open source project, the odds seem to be that it will either fail, or eventually be bought by a company that sees the project as an asset to its overall strategy.

The question is, how is this trend going to affect those who contribute to and/or use these corporate-sponsored projects? This is just one more piece of due diligence that companies need to do before depending on an open source solution. The answer (I think) is that it’s going to be an inconvenience, at worst, for those projects with strong enough communities to rally behind the project if the corporation fails (or is perceived to fail) in its stewardship of the project.

This trend isn’t that different from proprietary software players, except that it’s sort of new to the open source space– it’s still a novelty when an open source play is bought for hundreds of millions of dollars, while this sort of thing has been commonplace in the proprietary world for years.

It’s also different in that we, that is customers and users, have a fallback that the customers of proprietary software don’t have: We have the opportunity to fork the code if an open source company goes in a direction that we don’t like.

For example, think for a few moments about recent acquisitions in the proprietary software world, like Adobe’s purchase of Macromedia. Users of Macromedia Freehand got to see it discontinued when Adobe decided it didn’t need to continue developing a program that was so similar to its own Adobe Illustrator. I’m sure you can supply your own examples of proprietary packages that were discontinued, or just seriously bungled, after the company behind the software was purchased or merged with another software company.

Adobe wouldn’t have been able to do this, of course, if Freehand was an open source project. It might have faded into obscurity if Freehand lacked the community to keep it going– but at least users would have the option to keep the project alive.

The open source model, then, serves as an effective bulwark against a company buying a competitor to take it out of the market. It also provides an escape hatch for projects that are snapped up, continued, but mismanaged. That’s actually a pretty significant advantage for users and customers over the proprietary model.

The real question is, does a project have a community that can step in and take over development? While MySQL doesn’t have a particular large external development community when compared to a project like PostgreSQL, it’s pretty obvious that with as many users and organizations that depend on MySQL, it’d be able to muster support if it ever came to that. Can you say the same about the other projects you depend on? Food for thought as we see a year full of consolidation on the horizon.

I’ve often wondered why Apple maintains a server OS and product line. From what I can see, they don’t seem at all serious about the server market.

Just because I’m a Linux guy, it doesn’t mean that I don’t find other operating systems interesting. First and foremost, I’m a geek, and I always want to try out the new shiny. I particularly like to pay attention to what other *nix OSes have going on, and the features that will make life as a system administrator easier.

With the release of Leopard a few weeks ago, I was moved to peruse the features you could find in Leopard server. While I haven’t had the chance to get any hands-on time with Leopard server, the feature list looks interesting — particularly for SMBs that might not want to try to hire an admin.

But, looking at Apple’s product line and server features, I have to ask myself if these folks are really serious about competing in the server space.

First, there’s Apple’s policy of only allowing its OS to run on Apple-bought hardware. This is an extremely annoying policy for desktop users, and near lunacy for any company that’s at all serious about getting a reasonably sized slice of the server market pie.

The folks at Sun have come to realize this, and are reaching deals to offer Solaris on Dell systems. I’m not ready to switch to Solaris, either, but at least I’d give Solaris a second look because I have some flexibility with regards to buying a server to use Solaris.

Where are the blades?

I also wonder, where’s the variety? If you look to HP, Dell, IBM, Sun, et al, you’ll see a wide variety of hardware: Low-end 1U servers, four-CPU heavyweights, blade servers, and other assorted options. Most organizations should have little trouble finding systems that meet their needs, thanks to the wide variety of open systems options.

And then there’s Apple. Assuming I’m willing to buy hardware through Apple, what are my options? You have the 1U Xserve, which is a respectable 1U system that can handle up to 32GB of RAM (at a whopping $23K for eight sticks of 4GB RAM…), two quad-core Xeons up to 3.0GHz, and three drives, and you have the Xserve RAID for storage.

And that’s it. Two systems to choose from. No blades, no big iron, nada. I’d pass on Apple’s offerings based on its limited selection of hardware alone.

What about Virtualization?

The other thing that really keeps me from taking Apple very seriously in the server market is its lack of a virtualization strategy. At the risk of sounding like I’m accusing Apple of failing to be buzzword-compliant, I’m not just saying that virt is important because it’s a hot topic these days. I’m bringing it up because virtualization is a key component these days in most companies’ IT strategy.

Forget the buzzword for a moment, and answer me this: What’s Apple’s answer for server consolidation and resource management? Assuming an organization does want to deploy Mac OS X, how can they partition their hardware for best resource utilization?

I’m sure part of this goes back to Apple’s need to control everything. I’ve talked to folks at various virtualization vendors (say that three times fast…) and everything I’ve heard indicates that Apple has no plan to allow Mac OS X (server or desktop) run as a guest host under virtualization — even on Apple’s own hardware.

OK, fine. But why doesn’t Apple offer its own virtualization offering? If Mac OS X had something like OpenVZ or Solaris Zones, that might be good enough. (Assuming an organization wants to be all Apple which is not likely in most cases.)

Now, being Apple, it’s entirely possible the Cupertino crowd has an “insanely great” virt strategy they’re waiting to spring on the world at one of the developer conferences or whatever. For all I know, Steve Jobs will be showing off a way to manage virtual machines using nothing more than a Front Row remote sometime in early 2008, but the company hasn’t disclosed anything regarding virtualization so far.

Anybody Got a Roadmap?

Actually, that brings me to my final point — Apple holds its cards far too closely to the vest for the server market. When I speak to major IT vendors like IBM, HP, Dell, and Sun, I can at least get a picture of the short and long-term plans.

What’s Apple gonna do? They’re not telling. Again, this is a strategy that’s worked for the company on the consumer side, but it’s not working for the IT side of the house.

OK, I’ve slammed on Apple long enough, but the question remains: Is Apple serious about the server? If so, I think they need to change their strategy. Despite my Linux leanings, I think that Mac OS X could be a contender if Apple could get out of its overly proprietary mindset, at least on the server side.

No doubt you’ve heard the prediction before — “this is going to be the year of the Linux desktop.” At the risk of being repetitive, though, I’m going to go ahead and say it: 2008 really could be the year of the Linux desktop.

Yes, yes. I know — we’ve all heard this before. If I recall correctly, 2001 through 2007 have also been” the year of the Linux desktop,” according to various pundits. Hear me out, though, because it seems a few vendors are starting to get a clue about how they can make Linux compelling.

For quite some time now, I’ve been a bit skeptical about the chances of Linux landing on the desktop in any great numbers. Even the Dell deal with Ubuntu earlier this year did little to boost my confidence in desktop Linux, because the PC giant seemed to be responding to Linux desktop demand with much less than full enthusiasm. Dell’s offerings may be great value for Linux users who know they want Linux on a desktop or laptop, but the Ubuntu Dell systems are relegated to a dusty backroom of the Dell site that won’t be visited by many mainstream users.

So, the fact that Dell is offering Linux isn’t the Holy Grail that many Linux users thought it would be. The” if you build it, they will come,” philosophy just isn’t going to get it done for desktop Linux. Nor is technical excellence alone going to propel Linux into the desktop market. Vendors need to offer solutions that play to Linux’s strengths and put a little marketing muscle behind it as well.

Linux isn’t a direct Windows replacement, and users seeking a drop-in replacement for Windows often come away disappointed. Not because Linux is unusable as a desktop, but because it just isn’t Windows. A lot of people, though, don’t really need Windows. They just need a system that handles basic functions and doesn’t cost an arm and a leg.

I’m talking about the Asus EEE PC and the Everex Green gPC. Both systems are, by today’s standards, cheap, underpowered, and limited in function. They also play to Linux’s strengths by being cheap, underpowered, and limited in function. Despite limited horsepower, both systems are more than adequate for performing basic desktop functions (email, Web browsing, running OpenOffice.org, etc.) and offer stripped-down interfaces that are less confusing for new users.

They’re also, as of this writing, both sold out. The gPC has been selling through Wal-Mart, and sold out in just a matter of days. The EEE PC was sold through a number of online retailers, and also sold out in just a few days. I’m sure both systems were in limited supply, but there’s clearly pent-up demand.

So we have two desktop categories screaming for an alternative. Cheap PCs that will allow home users to do the basics without having to foot the bill for a Windows-capable PC plus the Windows licensing fees. And portable systems that can be used to complement a users’ existing desktop or laptop. The Asus EEE PC, with its tiny screen and cramped keyboard, isn’t going to be a full desktop system for most users (unless they have really tiny hands and enjoy squinting) but it’s perfect for travel. I wish I’d had a EEE PC on my recent trip to Supercomputing 2007 rather than the 15″ widescreen laptop that is just too bulky for the standard cattle-class airline seat.

Do I think that we’re going to see Linux devices overtake Windows in 2008? Not likely. But I think this is going to be the year you’re going to see a lot of companies offering slimmed-down devices boasting Linux for average users, which just might be the stepping stone for Linux to start encroaching on the full-fledged desktop market as well.