The open-source system, which features a modular architecture, is already dominating the smartphone market.
Seemingly, mobile operating systems have reached the point where they're all "good enough" to do the job. So, now, arguments regarding their benefits are shifting to areas such as openness as a way to stimulate competition and suppress costs.
Historically, mobile handset capabilities were closely tied to their hardware platforms. The first generation offered one- or two-line LED/LCD displays. Then came phones with text-based menus, with the ability to handle text messaging as well as voice. The 3G mobile phones offered simple graphical user interfaces (GUIs).
Today, we have phones with full multimedia GUIs that can handle text messaging, picture messaging, e-mail, Instant Messaging, video clips, etc. In addition to the now ubiquitous built-in camera, they can surf the Web and run advanced applications. Basically, the focus has shifted from hardware to software (Fig. 1).
However, while handset capabilities have become more sophisticated and more complex, mass-market handset makers find it increasingly difficult to keep pace with market demands. The reason has more to do with history and commercial interests than technical capabilities.
Designing and building mobile handsets has always been an arduous process, involving reams of documentation and months of interoperability testing. Each handset model is virtually built from scratch. The irony is that, despite all the groundwork, handsets often vary quite widely in the way they interpret specifications. As a result, operators run into problems in getting their services and applications to work the way they would like across all of the available handset models.
For example, it's taken years for MMS to achieve somewhat consistent operation across different handsets and networks.
The complexity of the definition process slows the speed of evolution and limits the ability to experiment, which is critical in the development of data services. One way handset makers have tried to get around this issue is to build multifunction devices that promise to be all things to all people—the Swiss Army knives of the mobile world. Often, though, these devices don't excel at everything, and users tend to use them for a limited set of functions only. Take Apple's iPod—its success comes from the ability to perform a small set of functions very well.
In today's competitive market, mobile operators want to be able to roll out innovative value-added services that tailor to different customer segments quickly and cost-effectively. They require new handsets that can make the most of the latest mobile-network capabilities they've invested in, and be easily customized to appeal to different target users. The trouble is, proprietary monolithic OS architectures present more of a barrier than a solution.
Linux enters the mobile fray
Many mobile operators, handset makers, and application developers recognize the huge potential of Linux and the ability of the global open-source community to meet its needs. As Fig. 2 shows, Linux is already taking significant market share in the smartphone market, growing at an annual rate of over 1000% between 2004 and 2005, says market analyst Gartner.
Linux is a modern operating system that's suitable for running complex, mission-critical applications, which is why it now dominates the server environment. Unlike most "homegrown" mobile operating systems, Linux features a modular architecture with loadable device drivers and modules.
This allows for abstraction from the underlying hardware and between components—a very important factor when building a complex software system. As a result, enablers can deliver a range of functions using, say, shared databases, digital rights management (DRM), and security features. In short, Linux lets manufacturers build more robust and secure handsets that offer more features and more predictable quality than with proprietary operating systems.
Being open source, Linux gives OEMs and ODMs control over their business destiny. With complete control over the software in their phones, manufacturers are free to create phones that the market wants—quickly and cost-effectively. They needn't depend on the business goals of proprietary OS providers.
One very attractive benefit of Linux for mobile operators is the opportunity for total customisation of the phone software and user interface to suit their own brand positioning and customer targeting.
The customer experience is strongly influenced by GUI design. The beauty of Linux-based platforms is that they allow for both differentiation and consistency of interface design and branding across different mobile-device types. What's more, Linux gives handset developers a single code base on which to develop a range smartphones and feature phones. Operators and users will also benefit from the wide availability of third-party applications coming from the large global Linux community and ecosystem.
Linux also cuts cost and time-to-market. With more flexible hardware and software, device manufacturers can better control costs. Linux drives commoditisation of the software stack, and provides more predictable software quality than proprietary systems (through constant testing and bug fixing), shortening time-to-market.
Building a complete Linux phone today would be just as expensive and time-consuming for an OEM as it would be to build a phone based on any other OS. However, integrated reference stacks (including everything from hardware to user interface) are on the way, giving developers a predictable platform on which to create phones that meet customer needs.
Benoit Schillings is chief technical officer of Trolltech.