OSCON, Pt. 2.3: Jim Zemlin's Outlook Is Cloudy (In A Good Way)OSCON, Pt. 2.3: Jim Zemlin's Outlook Is Cloudy (In A Good Way)

Wednesday morning at OSCON I sat down with Jim Zemlin of the Linux Foundation, who I'd also checked out at the mobile Linux panel on Monday and had sat down with at the Red Hat Summit earlier this year. He's a terrifically engaging talker, and regaled me with tidbits about his current Linux-oriented fascinations: cloud computing and the mobile Linux market.

He's convinced (as I am) that the netbook market is red-hot, not just in general but as a breeding ground for Linux in particular. What he's also fascinated by is the potential to use Linux to build extremely specific kinds of net-connected devices that have no existing, predefined niche.

"If I would start a business tomorrow," Jim mused, "I'd do it in the netbook marketplace. I'd build a dead-simple $200 device that targets sports fans, women over 40, people who want healthy lifestyles. This would be a small subnotebook with a shoe sensor by Nike (like the kind that integrates into the iPod), a heart-rate monitor, a Web back-end database for caloric intake, a custom Linux kernel, always-on Internet connectivity, and so on. For revenue you'd probably charge subscription fees for integrated services that are all about the same sorts of healthy-living things. Everyone wants to be thin and young!"

Cloud computing has, in his purview, the same game-changing potential as open source itself. "I recently talked to a big bank's CIO and he told me their cost per CPU core-hour was something like thirty-five cents. Amazon's cost per CPU core-hour is about seven cents. Latency is still a big issue with clouds -- obviously you don't want to be running a trading application on one of them -- but before long it's going to be very hard justifying not using a cloud for something. All these new startups use clouds and virtual machines to scale on demand; they don't buy any iron at all."

I pointed out that in my own experience, the problem with cloud computing isn't the metaphors used or the technology: it's figuring how much you owe.

"Yeah. One of the tricky things for these cloud computing infrastructures is billing -- computing chargebacks based on usage, and so on. At the company I used to work for, we could never figure out our own per-customer economics -- for instance, how much does it cost us to provide on a per-seat basis a PeopleSoft HRC? But it seems like this stuff is now getting figured out pretty well.

"You're definitely going to see more innovations with rate plans in cloud computing. Capped billing, flat billing, rates based on availability (you want four nines or five nines?), that kind of thing. Charging more money for more nines makes sense -- when you're driving a car and you go from 60 MPH to 70MPH, you're using twice as much gas. Going from four to five nines is a leap of an order of magnitude as well." Aside from selling access to the cloud, other opportunities for monetizing cloud computing might include management tools and backup/replication functionality.

(I mentioned that stuff like Amazon's recent service outages might cause trepidation for those just getting into cloud computing, but Jim described those as "growing pains".)

On the subject of Sun, Jim was less charitable: "I've been watching them and I think they're in real dire straits. Their stock is possibly at an all-time low, and they've lost a huge degree of critical mass in the marketplace. People now compare them to Digital [Equipment Corporation]. OpenSolaris and tools like Dtrace are not going to save them, and neither will ZFS. The smartest thing they could do is release ZFS in a Linux-compatible license, but I don't think they're going to do it. I don't think Schwartz is making bad moves per se, but once you lose that much critical momentum it's hard to catch up. The world has decided the server marketplace is a Linux and Windows world; this was over years ago. Embedded is an obvious area of high growth, and Linux is sewing that up. Mobile, too." Jim's words on Monday's mobile Linux panel echo the same sentiment.

A good chunk of the conversation at this point veered into a digression (albeit a fun one) about Microsoft's presence in the game-console market, but we got back on track when I mentioned how many consumer devices are now using Linux in a subtle way.

"I'm a gadget guy, so I try to get my hands on every Linux-based gadget (my wife hates me for this). My favorite is the Sonos music player. Linux kicks butt for this kind of stuff because it's totally license-free, totally malleable, totally brandable.

"I had lunch with the CEO of a major commercial distro the other day, and they were talking about how well they were doing -- they're a terrifically run company -- but their biggest competition is a free respin of their own product! But the interesting thing about that is, that's the GPL -- and they offer a much higher degree of value over a free alternative, and make money off it. That's the promise. You're competing at that much higher a level. It raises the bar for value provided through open source."