Free Newsletter
FEATURE: Mobile Linux: Why it will become the dominant mobile OS
By J. Gerry Purdy, Ph.D. VP & Chief Analyst for Mobile & Wireless at Frost & Sullivan.
In the world of PCs, there's one major OS--Windows and a couple of smaller operating systems including the Apple Macintosh OS X and Linux. In the server world, there's about an equal mix between Windows and various versions of Linux.
But in mobile today there are a plethora of embedded and open operating systems including Symbian, Palm (legacy), Windows Mobile, Linux (MontaVista and PalmSource), BREW and RIM as well as embedded systems from Enea and Wind River. Mobile Linux will become the dominant OS on mobile phones within five years. Here's why.
The current mobile OS solutions lock developers into a fixed architecture. This yields user experiences that have a homogeneous look and feel, but these environments make it difficult for developers to port their applications from one platform to another. Java doesn't help because there's simply too much difference between the devices. Up to 40 percent of some commercial phone application development engineering budgets are focused simply on porting to different mobile devices. I'll be writing about the porting process in the near future, but for now, I want to focus on the mobile OS and why Linux appears to me to be a good direction for the entire mobile market.
Take a look at the composition of the "software stack" made up of the hardware interface (device drivers), kernel (provides basic OS services such as task switching and file management), user interface and then applications.
In Windows PCs, the OS plays just about as important a role as the UI and applications. And, that's appropriate since services like task switching (running different applications) and file management and Windows Explorer are used just as much as the applications. But, in phones, the OS--particularly the kernel--is less important than the services, UI and the user experience. Thus, the value in mobile OS and applications has shifted from the OS being of equal value in PCs to the services and applications being at a much higher level than the mobile phone OS.
In mobile phones, the OS kernel is thought of as being embedded or "underneath the covers." The value has shifted from the OS to the services and applications. This creates an opportunity for the mobile OS to be "open source." Thus, Linux seems poised to sit perfectly well in the new mobile OS/services/apps ecosystem. The core or kernel becomes free and shared via open source with everyone providing their "value added services" to the core OS. The entire community contributes and the entire community benefits.
Linux was not originally designed for mobile but MontaVista and ACCESS (PalmSource) are working on power management, devices drivers and other services to enhance the basic Linux OS to make it "mobile friendly."
Mobile devices developed under an open source operating system have a distinct look and feel that reflects the developer's R&D investment, without the licensing restrictions that can stifle creativity and blur brand identity. Linux has the support of hundreds of thousands of developers. Linux also makes cross platform development to a number of (typically ARM-based) microprocessors easier and reduces the complexity of porting.
The OS kernel is difficult and costly to produce (multi-tasking, multi-threading, file system, interface management, etc.) but Linux has been developed over many years to provide excellent kernel services at low cost. In mobile, it's more important to work on the user interface and user experience than it is to focus on the OS like it is done on the PC.
PalmSource acquired China MobileSoft in 2004 to gain China MobileSoft's optimized Linux OS and applications that were already running on phones in China. PalmSource was itself then acquired by ACCESS (HQ in Japan) in 2005. The ACCESS Linux Platform (ALP) is a set of software utilities and services running on top of a Linux core provided by open source or from firms like MontaVista. ACCESS promises to provide backwards compatibility for Palm OS applications. That will enable them to leverage over 400,000 developers and bring them into the Linux mobile ecosystem.
Motorola began Linux development in 2001 and by 2004, Motorola had a number of phones running Linux through a partnership with MontaVista. Nokia's Linux operating system, called Maemo, is based on Debian Linux. The Nokia 770 Internet tablet incorporates Maemo and provides Internet access, VoIP and WiFi. It could be the precursor to an interesting family of devices targeted for WiFi regions in major metropolitan areas which would rely on pervasive WiFi coverage to provide users with free voice (via GoogleTalk or Skype) and basic Internet access.
Samsung, Trolltech, NEC and Panasonic are all working on Linux-based devices and Texas Instruments has developed a mobile phone reference design for handset operators that uses Linux from MontaVista.
So, what about Microsoft, Symbian, BREW and RIM? Microsoft will still play in the enterprise. And, Symbian will argue that they have more experience in mobile development. But, if the big momentum carries forth with Linux, then Microsoft and Symbian will have to adapt their services to run on top of Linux. It seems plausible that UIQ (a Symbian subsidiary that focused on the user experience) should operate on top of Linux rather than the Symbian core to be more efficient. Economically, I believe all of the current mobile OS vendors will come to realize that the value in the total offering is in the services and applications not in the OS kernel and, as a result, will shift their offerings to become value added suppliers to the Linux ecosystem. Further, new companies like Funamobl offer open source applications so the community of developers can add new mobile devices to extend the reach of the applications.
As a result of these major shifts in the mobile ecosystem, we'll see more and more phones based on Linux come to market over the coming years. It's a wave that will be driven by lower cost. It's a wave driven by economics of the ecosystem. It seems like a wave that has too much momentum to not happen.
I see the mobile Linux wave out in the ocean and am sounding the alarm--the mobile Linux tide is rising. Be prepared.
Comments
Posted by Eugenia | January 3, 2007 - 5:04pm
I have replied in your article, in the form of an open letter, here.
Posted by stelt | January 3, 2007 - 8:22pm
There's a very nice common ground on all the different mobile phone operating systems, see http://svg.org/special/svg_phones
Posted by matt | January 4, 2007 - 2:22am
I'd love to light a candle and sing ""Kumbaya" with you around a table of companies sharing a linux based mobile operating system, but it's simply not going to play out like that. The Linux Phone Standard so far has produced nothing but a bunch of press releases saying how they are all going to work together (and they've been at it over a year) What is actually going to happen is motorola, access and a bunch of other companies will share the base of the os (the kernel, and drivers, maybe some low level libraries) but Moto, and each other company will build there own proprietary libraries on top of the kernel and you will have applications written specifically for palm o.s. motorola, samsung ect. Even if these companies wanted to share and play nice and develop an os they are YEARS!!! behind ms and symbian which is like light years when it comes to developing an o.s. + applications+ a user base... I think that this is going to be the year that Symbian makes it into the U.S. Market, or Microsoft takes over and dominates with Crossbow (when it comes out). I am hoping for the former...
Posted by smeck | January 4, 2007 - 4:56am
matt, that's a reprise of the arguments i heard against desktop linux five years ago.
enough said ;)
Posted by UMPC architect | January 4, 2007 - 9:46am
Excellent article and points in favor of Linux.
I will also add that Linux requires less HW resources (implies lower BOM cost, battery, weight,..). In addition, MS has stringent UI flexibility in their licensing.
Posted by Brian Empey, P.Eng. | January 4, 2007 - 2:00pm
Good high-level article. The fact is, Linux support is much better than most people think, and companies like mine (www.medallionsystem.com) have solved many of the same issues that the OSDL's Mobile Linux Initiative (http://www.osdl.org/lab_activities/mobile_linux) is tackling. And we aren't alone.
The other major FUD fuel from people who don't understand Linux at all (such as "Eugenia" above) is that Linux is Linux, and there is much more compatibility between the various offerings than there is between different versions of MS Windows! And you don't have to write to the pre-installed GUI. You can include your own GUI, since they all run on the FrameBuffer (FB), and that interface is standard on Linux. There is no reason to not include FLTK, GTK-FB or whatever with your application.
In addition, there is also a standard mobile phone GUI called OpenGL/ES (http://www.khronos.org/about/) whose members include some phone makers, such as Nokia, Motorola, Ericsson, Sony, SK, ... you may have heard of these. OpenGL/ES specifies a minimum LCD size of 220x176 pixels, which the LCD makers and phone makers are already adhering to. As this standard becomes ubiquitous, you'll have a choice of writing to FB at the lowest level, writing to OpenGL/ES at a higher level, or just writing JAVA Apps using .AWT ... and that's where Linux really shines. CHOICE.
But the biggest force that will drive the adoption of Linux (and not mentioned in the article, although the author hints at it) is that the tide of adoption of Linux will reduce the development costs for APPs to a level that a company cannot competitively work on another OS, because their shrinking R&D budgets will limit them to the OS with the most complete and mature stack. That is a position Linux has already achieved and will not forfeit for decades. Anyone who thinks Linux isn't at that level yet is either ignorant of the current reality (some are still looking at Linux 2.4) or they're in denial (the Palm, MS, and Symbian boosters) ... or both :)
Posted by liilliil | January 17, 2007 - 7:13am
Dear Gerry Purdy! Many major software companies, such as a IBM, Google, Novell etc. works for Linux, but this platform still looks like a cheap & marginal. Year passes, but it's just a bunch of crap. Nobody whats Linux on mobile pc. Noone, Gerry
Posted by liilliil | January 17, 2007 - 7:19am
Dear Gerry Purdy! Many major software companies, such as a IBM, Google, Novell etc. works for Linux, but this platform still looks like a cheap & marginal. Year passes, but it's just a bunch of cr@p. Nobody whats Linux on mobile pc. Noone, Gerry
Posted by Manoj | January 25, 2007 - 12:00am
My goodness, i for one would think twice before having linux on my phone. It's like a dark community i have always been so afraid to enter.
Besides, we will rather have developers to spend more time and money on porting rather than have the unnecessary overhead of a platform independent OS slowing down our mobile devices. These operating systems have to run as fast as they can on every mobile phone.
Posted by Eli Dickinson | January 25, 2007 - 9:02am
I don't see any reason to assume that Linux will be any slower than any other OS.
It's not really all that different from Windows Mobile (which runs on a variety of different architectures), is it?
Posted by Anonymous | February 8, 2007 - 5:46pm
In fact, Linux can be made to fit into a sub 8MB phone if you choose the componenets wisely. No other smartphone OS (Windows, Symbian) can even dream about it
Posted by haiqu | April 7, 2007 - 10:48pm
Good on you Brian Empey, for solving many of the OSDL's issues. We all look forward to this data being posted to the public domain.
As to Eugenia being ignorant of Linux ... well, at least I've heard of her as the editor of OS News.
[Bart Simpson voice ON]
But who the hell are you, maaaan?
[OFF]
We all remember the "success" of OpenGL at the PC level. Fact is that inefficient, compromise solutions are rarely used in production. That will include anything to do with Java for now, until the power in your pocket approaches the level of a PDA and can run for 24 hours between charges.
If Linux is to succeed in mobile phones it will do so due to choice of platform (as distinct from being locked into ARM-derived hardware) but porting of applications will, as Eugenia states, be impossible due to proprietary solutions to all else.
Then again, naybe we should stop looking at every piece of hardware that comes along as something to hack at and just use it as the appliance it was intended to be, no?
As in the day of BeOS (it was a Tuesday, I believe) these devices will sink or swim on the availability of a "killer app" that runs on them. Anyone can get a basic phone for peanuts, but if my Forex client will only run on Windows-based machines then guess what I'll be buying.
Posted by Movie | August 18, 2007 - 3:45pm
I have replied in your article too.
Sig:
Download films movies
Posted by oyun | October 10, 2007 - 4:25am
With the name Vista, we want this campaign to have a strong visual element.
http://www.googleoyun.net
Posted by watch_movies | October 14, 2007 - 9:08am
Very nice and well written guide. It's very helpful to me, Thanks
---
download movie
Posted by Anonymous | October 22, 2007 - 7:07am
I dont really think that Linux will be able to beat monstrous WM.
Yes, in home computer - it is really good.
On my PDA... NO THANX.
Posted by DrNiko | October 22, 2007 - 7:14am
I really think that WN on mobile devices is quite slow, but much more comfortable than Lunix-Unix.
-----------------
My Best Dvd Online Films Library.
Posted by Radyo Dinle | October 23, 2007 - 12:44pm
Mobilinux's basis in open source technology will save customers money. Additionally, the company says it will save cost by supporting single-chipset phone designs. Real-time determinism and the added performance afforded by modern compilers are central to Mobilinux's scalability, according to Lehrbaum, and will enable the embedded OS to move downmarket into phones that handle both voice and application processing on a single chipset, running a single OS, with a single bank of memory (and perhaps an external modem and DSP coprocessor).
Posted by Viktor | November 10, 2007 - 12:32pm
It is completely agreeable with the contributor (Gerry Purdy) the very cognitive and interesting article
Posted by bobic | December 7, 2007 - 5:49am
It completely agree with the author about mobile Linux. Also I consider that it will be dominating OS in mobile phones.
---
replica watch
Posted by Brian Empey, P.Eng. | March 11, 2008 - 5:24pm
RE: Available code
Just buy a dev.kit, and you get Linux, the source, the tools, and the source for the tools.
(all based on standard ARM-Linux)
WHO AM I?
Just a guy with a company that designs and builds award winning products that have other people's names on them. Our products help ensure wireless text messages move reliably, credit-card processing servers don't fail, patients get the correct medications, soldiers have reliable communications, and you have safe drinking water.
But you'll never hear of us.
Posted by bobic | March 18, 2008 - 3:16am
For what? We've known Palm was developing some kind of next generation platform for some time. I even mentioned it just a few days ago. The only unanswered question was whether it would be a totally unique architecture or just Garnet slapped on top of Linux. We still don't know what the hell this OS really is, other than the fact Linux is its foundation.
The best news to come out of this announcement today is that Palm isn't moving to Windows Mobile. Had Palm chosen this path, it would killed what makes Palm unique.
You can bet money the new OS will not be made available to existing users (who have long endured and suffered from years of FrankenGarnet), which means we all have to upgrade to new smartphones in order to access PalmLinux (whatever it's called). And it's much too late for that because my tax dollars are going to iPhone, just like millions of other users.
In a nutshell... this is too little, too late. This move should have come two years ago. Palm has annihilated its brand name and reputation. It will take every resource to get back what they've lost, let alone grow beyond.
---
buy premarin
Posted by bobic | March 20, 2008 - 5:53am
However, there is nothing to suggest that it will become the only OS or that it will even become as dominant as Windows once was. The early PC market resembled the early mass market for cars; you can have any OS you want, as long as its Windows. The truth is that there can be multiple OSes but only one commodity OS, and Linux will eventually become that OS. The server OS market is different. Much of the strength of Windows in the server market is the apps (MS Exchange, MS SQL Server, particularly). Linux is fast becoming the commodity OS.avodart
Posted by Thomas IJ | March 29, 2008 - 1:04am
I believe different OSes will appeal to different people. There are those who prefer to experiment with linux and perform their own mods while others just stick to what they've known because they view linux as a bunch of complexities. Consider Africa which is now rapidly adopting linux OSes like Ubuntu becoz windows has been bothering them with policies and licenses etc. It all depends on region and customer requirements.
Posted by Anonymous | March 14, 2009 - 11:40am
I send this from a xv6800 from Verizon that duel boots wm6.1 and Androil [Linux]. running in Android. as yet Android does not provide all the features on this HTC device, if it did it would now only boot Android Linux. And HTC will not support an OS that their device is not originaly shipped with. Fortunatly there are so many developers and hackers who enjoy working with Linux without the constraints of Microsoft. This alone is likly to ensure that Linux plays a major role as prefered OS on handheld devices.
Posted by Oyunlar | April 7, 2009 - 3:53pm
Very nice and well written guide. It's very helpful to me, Thanks..
Click here to get the FierceWireless email newsletter for FREE!
Be the first to comment