In a sagging economy, doing business as usual is rarely prudent and usually disastrous. Companies often must change the course of strategies, or in some cases accelerate them.
After a spate of business meetings during the marathon that is Mobile World Congress, one takeaway is the acceleration in the mobile handset market of operators moving to open source software-based devices. Two specific device types are driving the increased adoption of open source software stacks: smartphones and netbooks, with many operators and OxMs also forming plans for Linux-based mobile Internet devices (MIDs) in the coming year.
Projections for Growth
The trend is more than anecdotal. In-Stat notes that “smartphones with Linux OS (including Android) will see the highest growth and the second highest volume behind Symbian. Linux OS will outpace Windows Mobile, RIM and iPhone OS X.”
Meanwhile, industry analysts’ projections for netbook (running either Linux or variants of Windows) sales growth rates in 2009 range from 50 percent to 128 percent. Many of these are being offered through operators with (profitable) data plans underwriting their already low costs. MIDs are also starting to appear on the horizon, fitting squarely between the smartphone and netbook.
Indeed, operators are viewing Linux as a strategic terminal platform on which they can build new revenue sources, new business models and a new customer base, regardless of device type. A Linux-based terminal platform offers them significant benefits over the legacy and proprietary platforms that formed the foundations of yesterday’s mobile devices. Based on an open source license, these platforms give operators a significantly greater ability to customize, tailor and brand the platform for their network. And, as the base for various open source consortia including the LiMo Foundation, Open Handset Alliance and Moblin, these platforms offer operators the power to influence their development and direction. Many operators are reviewing where to put their investments, some are outlining plans for a single platform while others continue to work with multiple platforms.
Stacking the Deck
Operators, OEMs and semiconductor manufacturers are now taking these open platforms and creating reference designs that will accelerate product development. Operators are integrating signature applications and other legacy assets with the open source platforms to create an operator-specific version that will be provided to OxMs as the starting point for final product development. OxMs are rapidly building competency in these open platforms to be better prepared to quickly deliver mobile devices for the operators. Semiconductor manufactures are also investing to pre-optimize their hardware platforms with these open source stacks to reduce early development cycles, differentiate their hardware platforms and accelerate hardware adoption.
This cross-value chain collaboration points to a new form of industry cooperation enabled by open source. By starting with a baseline open source software stack and a pre-optimized hardware-software platform, operators and their suppliers are innovating on new mobile devices at more effective levels — the much vaunted value line shift higher up the stack that has long been promised by open source but rarely realized. By not having to worry about first enabling a video stream or building support for 14 Bluetooth profiles, when utilizing the community’s stack investments, the ecosystem can now focus on developing new consumer features, services and advanced user experiences that will drive chip, device and subscriber sales.
Getting to the point where all the benefits are realized is proving to be more of a challenge than some have anticipated for several reasons:
- Open source mobile software stacks are relatively new, and expertise is not as widely available as with older, proprietary stacks (but is rapidly improving). Indeed, who will develop these reference stacks is a major question for operators — are they in the software development business themselves?
- Open source license provisions and how they affect open source software’s integration with an operator’s signature applications must also be addressed.
- Discovering ways to speed the testing process for the integrated stack without sacrificing quality may be the key to achieving development project reduction goals.
- The specter of fragmentation haunts the reference stack architecture, as no one wants to find themselves alone on a branch two years from now with no inexpensive way to rejoin the mainstream.
Despite these issues, operators are clearly moving full steam ahead with a variety of open source projects, including Android, LiMo and Moblin. The opportunity to quickly get a differentiated mobile device to market that will drive average revenue per user and subscriber growth is too alluring to ignore. Operators are driving the ecosystem to make it happen — look for the mobile devices soon!
Jason Whitmire is general manager of mobile for open source mobile software developer Wind River.
I enjoyed the article. I note that the statistics from various sources agree in indicating Linux and open source for mobile embedded development is approximately 20% of the market and growing fast.
I feel the evidence shows the benefits are being realized for open source on netbooks/mobile embedded hardware. Steve Balmer’s comments agree – see http://fosslc.org/drupal/node/345
I’m not sure I agree with your reasons for the benefits of open source not being realized. Here are a few counter points/ thoughts.
1) Open source is great for learning, and developing community/ecosystems e.g. Eclipse, Mozilla, OSGeo, Linux, Drupal. I suspect any shortage of skilled workers is temporary.
2) Agreed about licensing concerns, though with Qt’s new license and other libraries – this is less of an issue. In truth, understanding your license rights and obligations is not unique to open source.
3) Testing is not a unique challenge to open source/closed source. Neither is testing for variations of a platform (think 2000, XP, Vista, etc.)
4) Is fragmentation really an issue – see http://fosslc.org/drupal/node/361