Forget Symbian! The “war of the OSes” in mobile phones looks rosy for Linux and Windows, according to a detailed, insightful, freely available whitepaper from ARCchart. Funded by Trolltech, the 27-page “independent” report overviews the entire cellphone software market, offering in-depth reviews of 15 software stacks.

Entitled “Mobile Operating Systems: the New Generation,” the ARCchart paper is emailed in PDF format to registrants. It was written by ARCchart Research Director Matt Lewis, PhD, along with Andreas Constantinou, PhD, a lead analyst with VisionMobile.

The paper aims to help companies and developers understand the “roles, functionality, lines of partnership, and competition across software products” within the market for mobile phone software — a “complex endeavor even for seasoned industry observers,” the authors note (LinuxDevices.com's editors second that!).

The paper begins by identifying the software components typically found in the software stack of mid-range and high-end cellphones. These range from “low-level” to “high-level,” as follows:

1. Kernel
2. Middleware
3. Application execution environment
4. User interface framework
5. Application suite

The paper notes that three years ago, Symbian and Windows were seen as the future, due to their “open” platform nature. Today, “flexibility” has replaced “open” as the top requirement, however, because operators and handset vendors want customizability, along with uniform, consistent interfaces and functionality across models and product generations.

Flexibility is certainly a Linux strength, but the open-source OS still faces some considerable challenges, the paper observes.

About 12 million Linux handsets have shipped so far, according to a Gartner study cited in the paper (other estimates, such as those from MontaVista, are much higher). However, Linux phones have not been marketed in the U.S., while only two models have reached Western Europe. “Linux distributions lack the feature set requested by operated in Europe and the U.S.,” the paper suggests.

Today, Linux implementations are differentiated “too far down the software stack,” meaning that there are significant implementation differences at low levels such as the kernel and middleware layers. This results in poor if any interoperability among existing Linux implementations, the paper points out.

Today, vendors such as Motorola maintain their own Linux distributions. Others piece together Linux implementations from multiple third-party suppliers. Additionally, “complete” third-party Linux phone distributions were announced or launched early in 2006, by vendors A la Mobile (story), Access/PalmSource (story), Purple Labs (story), and Aplix (story).

The paper duly notes Linux phone standardization efforts, including LiPS (story), the OSDL's MLI (story), and CELF (story). It suggests that “of greatest interest” could be the “Open Platform Initiative” (OPI) recently announced (story) by Vodafone, NTT DoCoMo, NEC, Panasonic, Motorola, and Samsung — a group that between them has shipped far and away more Linux phones than all other phone vendors combined. According to the paper, OPI intends to “publish APIs, architecture, and test suites aimed at helping adopters assess and demonstrate product conformance to the platform specification.”

The paper next delves into 15 phone software platforms, describing the components found in each. The numbers in the following alphabetically arranged list refer to the components in the list above.

• A la Mobile — 1-5
• Access Linux Platform — 1-5
• Adope Flash Lite — 3
• GTK — 4
• mini-GUI — 4
• MIZI — 2-5
• MontaVista Mobilinux — 1-2
• Nokia S60 — 2-5
• Obigo — semi-complete 2 and semi-complete 5
• OpenWave MIDAS — 3, and semi-complete 2, 4, and 5
• Qualcomm BREW — 3, and semi-complete 2
• SavaJe OS — 6
• Symbian OS — 1, and semi-complete 2 and 3
• Trolltech Qtopia — 2-5
• Windows Mobile — 1-5

The paper goes on to review the technology and roadmap of each stack above, before wrapping up with predictions and forecasts.

Conclusions

The authors predict that the “sales value line” — the point in the stack below which per-unit royalties can not be extracted — will continue to rise, as the kernel and middleware become increasingly standardized and commoditized.

Even today, kernel vendor MontaVista — whose software has shipped in some 25 million phones, according to MontaVista claims — receives only per-seat developer tools licensing fees, rather than per-unit royalties, despite the company's brief effort to sign up “shared success” licensees.

Another prediction made by the paper is that over the next three to four years, the phone software market will contract vertically. That is, stacks and offerings will become more comprehensive, through product development up or down the stack, partnerships, and acquisitions.

In general, the paper authors believe that Linux will “prevail over many of today's creditable contestants.” They expect top phone vendor Nokia to eventually port its S60 software to Linux, an expensive proposition, but no more so than paying $100M to license Symbian each year. They write, “The question is not whether S60 … will migrate over time towards Linux, but when.”

The authors also suggest that “Qualcomm's commitment to BREW will be questioned.” Qualcomm already relies on Windows Mobile for its European products, the paper notes. Additionally, Qualcomm now distributes Linux with some of its phone chipsets.

The paper ultimately concludes that 2006 will be the “turning point” for Linux as a phone OS. “The future for Linux and Microsoft on mobile devices is looking particularly rosy.”

The complete, highly detailed, and comprehensive 27-page whitepaper can be obtained via email at no charge by registering here.

 
This article was originally published on LinuxDevices.com and has been donated to the open source community by QuinStreet Inc. Please visit LinuxToday.com for up-to-date news and articles about Linux and open source.