Linux Device Roundup

Four Linux device experts offer their opinions on the state of Linux devices and tell you about their must-have favorites.
Are you optimistic for the future of Linux-based devices?

Kingman: Yes. If they are running Linux, they won't crash, and the battery will last a long time.

Lehrbaum: Yes, very much so. Silicon vendors now favor Linux as the number-one platform to get their new device-oriented processors, chipsets and peripheral controllers up and running on, so Linux support gets a strong head start and is generally promoted by chip makers.

Powers: Oh, without a doubt. In fact, although I was joking a bit regarding “taking over the world”—I think Linux will continue to spread into the embedded market. It just makes sense. Also, with projects like Moblin and its ilk, embedded Linux on devices is looking really snazzy.

Weinberg: I remain very bullish on Linux as embedded systems software, notwithstanding announcements from Nokia, Symbian and others (see my blog, address in the Resources for this article) and advances by Microsoft. These and other moves/gestures toward openness and FOSS “scratch the itch” for access to source code as documentation and for source escrow, but they don't offer the unique combination of community-driven development, scalability, performance and real self-determination that you get with Linux and accompanying FOSS.

Do you consider any particular devices from the past year or so to be game-changers for the success of Linux? If so, why?

Kingman: The Wind River Linux Platform for Infotainment, for showing Linux could crack the automotive OEM equipment market; Motorola's Rokr Z6, for showing that a Linux phone could ship in volume in the US; the Netflix Player, for showing how inexpensive and powerful Linux multimedia devices can be; the myriad Orion-based NAS devices, for making NAS affordable to home users; and low-cost, power-efficient Nettop and Netbook devices, such as the Eee PC, for bringing desktop Linux to the masses.

Lehrbaum: I really like the Netflix movie-streaming set-top-box (manufactured by Roku). Linux has long been a winner in TV set-top boxes (think TiVo), and it's an area of exploding interest, given the growing ubiquity of broadband and drive toward streaming content to everyone's home theaters.

Automotive infotainment systems—featuring GPS, traffic updates, Internet access, streaming media, VoIP and so on—is another area set to explode.

There also are two+ major emerging device categories that both typically either come standard Linux or offering Linux as a full-fledged alternative to Windows: MIDs (mobile Internet devices), Netbooks and Nettops. All these terms were coined by Intel. MIDs strongly favor Linux due to being more appliance-like with built-in applications and not a lot of capability for normal users to alter the application set. Netbooks, typified by the Eee PC, have fully functional OSes, albeit stripped to fit in limited resources (often Flash, though HDDs sometimes are available as an option). They are generally available with Linux-only offered for the least expensive models and a choice of Linux or Windows XP Pro for the higher-end models (which have more RAM and storage Flash or HDD). A variant of the Netbook is the Nettop, having similar computing resources (including chipsets) but packaged in a mini PC-style box rather than the mini laptop-style formats of Netbooks.

The [articles on Netbooks by Lehrbaum, linked to in the Resources section] project the Netbook market will reach 50 million units by 2012, up from about 5 million this year. Obviously, given Linux as a baseline OS in the low-end models, this could be good news for Linux. However, as costs come down and RAM/Flash becomes higher density, the barriers to using Windows (depending on Microsoft's price positioning) could make Windows affordable. But clearly, the lower the end-user pricing of a Netbook, the less likely it is going to be able to afford a full-function MS OS like Windows XP or Vista.

So there is exciting potential here for Linux. Furthermore, bearing in mind that the whole idea of Netbook is the Net—that is, Web-based applications are central to its functioning. Thus, a Netbook is a bit like an Internet-connected thin client, and Linux does very well in such scenarios. Consequently, Netbooks should be considered a very high-priority target for Linux, just as mid- to high-end mobile phones are an important battleground for embedded Linux.

And, just as Netbooks are a fertile field for Linux, the same is true with Nettops—devices in which much of the heavy-lifting apps are Internet-based, and the device itself mainly needs a browser, e-mail client (not even required), media players and other basic functions, but are not expected to be true PCs that run every app you might want to try to load from a DVD.

Powers: Yes, indeed. The Eee PC began a trend that not only caught on like wildfire, but also significantly displayed Linux as a viable operating system for standard computer usage. We're just beginning to see how the Linux Netbook idea will change computing. The Netbooks are smaller than standard computers (or even notebooks), so they have a lot in common with handheld devices, and yet they are fully functional, so they demonstrate some of the same characteristics as a standard desktop solution. I think Netbooks might bridge the gap and open the door for vendors to take another look at pre-installing Linux on OEM hardware—even on the big desktop machines. That's my hope anyway.

Weinberg: Embedded Linux is already incredibly ubiquitous in intelligent devices. The real question is “What would change the game?” I think there are two vectors that could boost embedded Linux positioning:

  1. Truly open mass-market devices running Linux plus enabling middleware that would engender and excite both ISVs and a targeted developer community.

  2. Highly differentiated devices where Linux at the core would make a real impression on end users and build brand equity.

I haven't seen either of those situations emerge yet, but then again, other embedded platforms don't enjoy either scenario. Most RTOSes are 100% invisible to end users (except when they fail). Even Windows Mobile does not enjoy ubiquitous end-user pull, nor much popularity among developers, even if in some markets it's the only game in town.


James Gray is Products Editor for Linux Journal.