Linux crashes the mobile party

Cutting costs by deploying Linux is a well-established strategy on the server and even the desktop, but what effect could it have on the cost of mobile computing?
Written by David Meyer, Contributor

Very few people lie awake at night fretting over their choice of mobile operating system. In fact, very few people even know what operating system their handset uses.

Yet, despite this (probably healthy) level of ignorance, a quiet revolution is taking place in the mobile industry. Linux, the platform of choice for servers and geekier desktops the world over, is slowly winding its way onto the high-end smartphone, although no-one is really certain what it will look like when it gets there.

Part of the problem — although some would no doubt view it as a strength — is the number of organisations pushing mobile Linux. There are two main industry groupings dedicated to the cause: the Linux Mobile (LiMo) Foundation and the Linux Phone Standards (LiPS) Forum. Some companies are members of both. Others belong to one of the groups, but are implementing their own flavours of mobile Linux. It is little wonder that mobile Linux's chances of making inroads in the enterprise space have been played down by some very vocal critics.

Symbian, which does have a foothold in enterprise, thanks largely to its work with Nokia, took aim at the open-source operating system in July, when its head of enterprise business market development, Andrew Moran, dubbed mobile Linux "fragmentation city" and claimed it was "completely unfeasible" for business use. But then, Symbian is hardly likely to welcome Linux with open arms.

Gartner analyst Nick Jones has a more objective stance but he agrees that the platform is not yet consistent or standardised enough to be a serious proposition. "I would advise IT managers not to have anything to do with mobile Linux at this point in time," says Jones. "Imagine I'm an IT manager contemplating standardising on a mobile platform. I want something rich enough to deliver applications, that's available from multiple manufacturers, offering a decent range of handsets with corporate features. Linux just falls down on all of those."

However, those who have thrown their weight behind the movement are confident that it will succeed in an analogous way to Windows Mobile. Bill Weinberg, general manager of business development at the LiPS Forum, certainly sees Microsoft's mobile play as the best point of comparison for mobile Linux as a development platform, although he suggests that "Linux does it with a single code base end to end, not five or six distinct code bases with their own histories and bugs".

"Microsoft has a dominant space because it is very easy to extend corporate applications onto mobiles using Windows Mobile, but we're seeing Linux increasingly adopted as the operating system for use in corporate environments," says Adam Lawson, product director at Trolltech, the company behind the popular open-source development platforms Qt and Qtopia. "We would expect to see that trend extend into the mobile space in good time."

Lawson points out that governments in countries such as Brazil and China have officially backed the use of Linux in the public sector, and touts the security of open source as...

...a reason for this broad adoption. "Linux from the word go has been designed as a multi-user operating system," Lawson says. "It is easier to restrict access to particular data files and capabilities and prevent malicious access by third-party applications — the raw materials for that are available in Linux."

Lawson also thinks that another key strength of Microsoft's — Windows' ubiquity and the leagues of developers writing for the platform — could in turn become a strength of the mobile Linux movement. "Microsoft skills are widely available, but Linux is increasingly taught in universities. There are generations of engineers coming out with the relevant skills," he says.

Trolltech's Qtopia application-development platform, which is built on Qt, is geared towards embedded devices and handsets — notably Skype's internet telephony devices. "IT managers should be thinking about VoIP," says Lawson. "Expect to see Linux and Qtopia appearing on VoIP phones and desk phones — you are unlikely to see Windows running on a desk phone any time soon."

"Linux on the desktop and the mobile puts pressure on Microsoft to push their prices down, which is also good for the IT manager," continues Lawson. "Within five years we will see widespread adoption of mobile Linux as part of the corporate infrastructure."

However, it is clear that mobile Linux will be most widely seen in the consumer sector. Probably the most successful deployment of a Linux-based handset thus far has been Motorola's A1200 "Ming" phone, a touchscreen device which was one of the most popular phones of 2006 in China and elsewhere. The Qtopia-based platform used on the Ming has since acquired a name — MOTOMAGX. Motorola officially launched the platform in August with the declaration that it would be on 60 percent of the company's handsets within a few years.

"What we found, when we looked at our ability to come to market with products more quickly, was that Linux was going to be an extremely strategic platform going forward," says Christy Wyatt, Motorola's vice president of software platforms and ecosystems. "We expect it come down into the feature phone space, not just the smartphone. It's getting into our more iconic, mainstream handsets."

Motorola's motivation in backing mobile Linux is, according to Wyatt, the company's focus on the phone's software as a critical differentiator. "If you believe that the software experience matters as much as the hardware, then Linux makes a lot of sense," she says. "It offers a rich ecosystem and a rich set of tools, and there is a lot of investment happening across the industry. When we're dealing with an open platform we can easily respond to our customers' requirements."

Despite mobile Linux's lack of a track record in the enterprise, Wyatt suggests that — when the time comes — IT managers will welcome it with open arms. "Most IT organisations are already very familiar with Linux and there are a lot of analogies and lessons learned in that segment that we can take into mobile: for example, security. If they knew their devices were running Linux, it would probably give them comfort."

Wyatt speaks glowingly about the work of the LiMo Foundation, which Motorola founded in 2006 alongside NEC, NTT DoCoMo, Panasonic, Samsung and Vodafone (although companies such as ARM, LG and Broadcom have since joined). "We've invested a lot in creating that forum and are very happy with the number of partners," she says. "The opportunity for these large manufacturers and distro providers to share a common code gives us a stable platform that is attractive to developers and helps us target them in a consistent way. With Linux we have an opportunity to hopefully engineer past some of the mistakes we made with other platforms, such as Java and native implementations."

However, despite the LiMo Foundation's stated aim of creating...

...a shared implementation — as opposed to the LiPS Forum's desire to create a unified standard for mobile Linux — Motorola is arguably creating its own native implementation in MOTOMAGX.

"If you're Motorola, there is no huge incentive to be absolutely the same and consistent with other versions of Linux because you want to differentiate your products," says Gartner's Jones, who characterises the mobile Linux movement as "multiple groups of people with different motives heading towards targets that are moving". "Handset manufacturers want to have the cool device, and the role of the platform in differentiating your device has been demonstrated recently by Apple [with the iPhone]. There are obviously different tensions working along, and they are not all working towards convergence."

There even seems to be serious disagreement as to where MOTOMAGX comes from. According to Morgan Gillis, chief executive of the LiMo Foundation, the platform is "based upon LiMo Foundation technology and aligned with LiMo standards".

"The approach will be that the LiMo Foundation provides the base platform and some operators and some handset makers may choose to extend upon the LiMo platform," Gillis adds.

The LiPS Forum's Weinberg claims otherwise. "There are no standards to base it on," he says. "LiMo hasn't issued any documents, so it is hard to base a toolkit on it."

Weinberg characterises the difference between the LiPS Forum and LiMo Foundation thus: "What LiPS is doing is creating an open standard for the delivery of services and applications on Linux-based phones. The process, while open, is a fairly traditional standardisation process, with a body of members introducing requirements to specification. There will be a compliance regime emerging next year for that standards base. Implementation will probably arise initially from the LiPS membership, but there is also likely to be an open-source toolbox or even wholesale open-source implementation."

Pointing out that the LiPS Forum has already released a set of APIs, Weinberg continues: "LiMo is a group of companies that have got together and decided to create an implementation, not a standard, based on de facto standardisation. They will share a code of implementation exclusively among themselves, and at most will share a documentation of a subset of APIs they have created."

But Gillis is confident that, when it arrives, the LiMo implementation will bear fruit. "The progress made thus far by the LiMo Foundation platform provides good evidence that it is a platform that is likely to be widely adopted within the mobile sector," he says. "That's based on the composition and commitment of the original founder group, and [the recently announced] new wave of members, including important new members such as LG. One expectation is that there will be further important membership announcements before the end of the year, and this will signal there is a very good probability of broad adoption."

"The two different groupings [LiPS and LiMo] saw the same need within the industry at more or less the same time and got started in slightly different ways. It is still very much in the early stages," continues Gillis, before suggesting that it would be "quite plausible for convergence to happen in the coming period".

"I have no specific timescale in mind, but the naturalness and logic of this is evident. Our aims and objectives are very similar, if not identical," Gillis adds, while denying that any discussions have taken place between the LiPS Forum and LiMo Foundation regarding a merger.

Weinberg told ZDNet.co.uk that the LiPS Forum would welcome...

...co-operation with the LiMo Foundation. "I can see all sorts of ways to work together — if we were to merge our efforts that would be fine," he says, pointing to the LiPS Forum's track record of collaboration. "Last year we announced co-operation with another important standards body, the Open Source Development Labs, which has since merged with the Free Standards Group and is now called the Linux Foundation. Rather than duplicate efforts, LiPS will build on top of that in the mobile services and applications arena."

Yet all the collaboration in the world would mean nothing without the involvement of the mobile operators. Happily for mobile Linux, some providers have seen a lot of potential down the open-source route (Jones suggests that they see a unified approach to the platform as a way to reduce the power of the handset manufacturers). Vodafone in particular has become an enthusiastic sponsor, announcing last year that it wanted to pare its range of mobile operating systems down to a triumvirate of Windows Mobile, Symbian/Series 60 and Linux.

"Vodafone chose Linux as one of its three terminal platforms because it wanted an open, competitive alternative to incumbent proprietary platforms, offering manufacturers greater freedom to innovate," a spokesperson tells ZDNet.co.uk. "The key advantages of Linux for Vodafone are that it enables manufacturers to share development costs, thereby enabling greater product flexibility at lower costs."

"Linux also supports open participation in contribution, development and test environments, providing the basis for continuous quality improvement," adds Vodafone's spokesperson. "Open source is also extremely good at providing basic assets at a lower cost while the flexible licensing frameworks encourage competition and innovation."

Vodafone's spokesperson shows a middling degree of satisfaction with the pace of progress ("Vodafone is actively engaged with several suppliers on Linux-based product development, so in this respect we are happy with the momentum"), but predicts that products "in the mid to high tier" will appear in the operator's lineup in 2008-2009.

Ultimately, it may not matter what operating system a smartphone is running on. In fact, it may not even matter whether the phone is classed as "enterprise grade".

As the iPhone has recently demonstrated, an increasing number of business applications, such as NetSuite and WebEx, can work on any phone that has a browser. "Don't look upon mobile Linux as a platform to write applications to at this point in time — it is currently a platform on which to run Java or mobile web applications," suggests Jones. "There is a strong trend of web technologies driving into the mobile space. In a sense, the cellular platform is becoming indistinguishable from a slow broadband link."

"At the thick client end of the scale, you worry about the APIs, having a lightweight database on the device, salesforce automation, CRM — there you do worry about the operating system on the platform," says Jones. "But, increasingly, there are the less critical applications, where you are operating within signal coverage, where the web model is going to be good enough."

Editorial standards