December 22, 2005

Does mobile Linux really have legs?

Author: Jay Lyman

There is little disagreement about the opportunities for Linux on mobile phones, but it will take some work to give Linux its legs. While Linux today is one player among the many operating systems in the mobile handset market, finding its stride will be a matter of operating system enhancements, the formation of standards by the Linux Phone Standards Forum (LiPS), and use of the open source platform by a major manufacturer with a winning product, analysts said.

Yankee Group senior analyst John Jackson said Linux already commands the attention of the mobile handset industry, with platforms such as Monta Vista's mobile Linux operating system shipping on more than eight million phones in the last two years. "I think that Linux is of tremendous interest to the mobile supply-side handset vendors, platform vendors, et cetera," he said. "There exist stable, scalable Linux operating systems in the marketplace."

However, while Jackson mentioned mobile Linux supporters such as Motorola, NEC, and Panasonic, he also referred to the biggest hangup for Linux: the high cost of customizing and enhancing Linux for mobile handsets.

"The issue with Linux is Linux wasn't built to do anything," he said. "It needs to be optimized, and optimization is significantly expensive. It has tremendous upside, but the caveat is you have to make a significant investment in making it work, and that can't be underestimated."

Gartner vice president of mobile computing Ken Dulaney agreed, saying Linux was largely limited to dedicated consumer devices, and had less opportunity in the enterprise communications market.

"[Linux] is going to be a player, but the more popular choice is one of the proprietary operating systems," he said. "Once you get one of these tuned for wireless, you don't want to muck with it because it's hard to do."

Not so smart phone niche

Ovum wireless analyst Tony Cripps agreed that customization was becoming prohibitively expensive, but viewed it as an opportunity for Linux.

"My belief is that Linux could become a highly significant platform for mobile phones, but that it is more likely to show its greatest value in mass-market handsets rather than in so-called smart phones," he said. "The biggest problem for handset manufacturers is not smart phones. These are high-end, expensive devices representing a small part of the overall market and which are already well catered for by established operating systems/platforms such as Symbian, Windows Mobile, BlackBerry, and to a lesser extent Palm OS, as well as Linux-based platforms such as Qtopia from Trolltech."

Cripps said that the bigger problem for handset makers is adapting the often proprietary operating systems they use for their mass market phones to today's handset requirements. He said while Nucleas, OSE ENEA, itron, and others, as well as customized OSs from Motorola and Nokia, have proved stable and reliable for years, the changing hardware and functionality demands of even mid-range handsets is pushing these platforms to the breaking point because of their lack of flexibility. "Each must essentially be custom-built for each new handset model, and this is becoming prohibitively expensive and time consuming," Cripps said.

While Cripps included Microsoft's Windows Mobile among "more flexible, open" operating systems that may have opportunity here, he said the biggest chances for mobile Linux are in the mid-range mass market, where cost is key.

"[Windows and Symbian] remain expensive and will not deeply impact the mass market for some time," he said. "Linux's real potential appears to lie in its potential to replace proprietary OSs for mid-range handsets, where keeping the so-called software bill of materials (BOM) to a minimum is crucial. Linux's open source credentials, even in its mobile form, and the resulting adaptability to different hardware that comes from the efforts of the open source community will potentially deliver both the low cost and flexibility that handset manufacturers need to replace their increasingly expensive proprietary OSs."

Cripps warned, however, that such a scenario would produce mobile phones that may be as closed as today's mass market phones in terms of end users' ability to install new native Linux applications. "They will instead rely, as users do today, on Java virtual machine as the extensible application platform for phones such as this," he said.

More obstacles to overcome

Gartner's Dulaney said that the disparity among different uses of Linux may hinder its mobile growth. "It's a modern operating system, but in many aspects it's like DOS in the early '90s. The incompatibility from vendor to vendor makes it harder on developers. It's highly variable from manufacturer to manufacturer, and it does not resemble Symbian, or Microsoft, or anything. The problem with Linux is people tend to think it's one thing. It's not."

For his part, Ovum's Cripps said the biggest obstacle to more mobile Linux is its origins on the desktop and server, where computing and electrical power are readily available. Still, he said efforts such as OSDL's and the LiPS Form may help mobile Linux find its stride.

"Mobile phones put far more of a premium on things such as memory, battery life, and real-time performance," he said. "OSDL, in particular, is clear on this point and does not consider mobile Linux to yet be ready for mass market phones in terms of power management and response latency for real-time events, even in the most highly developed mobile Linuxes. Bodies such as OSDL are important to the development of mobile Linux both in terms of raising the profile of the technology and in terms of focusing the efforts of the community in achieving some well specified goals.

"OSDL's focus on mobile Linux kernel issues and hardware integration is well complemented by the efforts of the newly announced LiPS Forum, which is concentrating on higher layers of the Linux software stack. These include common application and service enablers -- this work is intended to be applicable across both fixed and mobile phones. Third-party application developers should then, in theory, be able to create differentiated applications around common foundations."

Cripps added what mobile Linux truly needs to succeed in the space is a major effort by a handset manufacturer to kick-start it, something that may be more likely to come from Asia.

"Motorola has stated plans that will possibly see it across its entire handset range, but it's not rushing to make this happen," he said. "NEC and Panasonic have built a few Linux-powered handsets for NTT DoCoMo in Japan as well, but perhaps the strongest near-term potential may lie in the Chinese market, where equipment companies like Datang see handsets, and especially ones with low BOM and little outside control with operators, as offering much higher margins than switches; and where operators like China Unicom are taking a great interest in handsets in their own right, much for the same reasons."

Click Here!