Intel's Wind River Hypervisor and technologies like it could usher in a new age of Device-agnostic Smartphone Operating Systems and unprecedented customer choice.
My editor-in-chief and colleague Larry Dignan this morning has outlined a number of the strategic problems Apple may have when moving forward with its current litigation with HTC.
Indeed, Apple may have bitten off more than it can chew if Google and Microsoft come into the equation. But I've found something else that may kill this lawsuit and others like it dead -- Embedded Virtualization technology.
Click on the "Read the rest of this entry" link below for more.
Virtualization? Isn't that what big companies use in datacenters? What the heck does this have to do with smartphones and the HTC lawsuit? Indulge me.
In 2008 I wrote an article that didn't get a lot of pageviews -- "I want an iPhoneStormDroid".
It was about Wind River's new virtualization technology that allows embedded operating systems to be abstracted from the hardware that it is running on.
Back then, Wind River was an independent, private software company that sold niche embedded software development suites for vertical market applications. In June of last year, Intel bought them for over $880 million. Recently, Wind River introduced a software development suite for building and optimizing Android phones.
Also Read: Wind River, Tasty Embedded Linux TreatIn my fairy-tale scenario -- or at least in what seems like fantasy today, I described a mobile computing world where a customer walks into a wireless carrier store, picks a "body" or device form factor that they want, activates service, then and runs whatever smartphone OS they want on it -- Android, Windows 7 Phone Series, Palm WebOS, Symbian, MeeGo, BlackBerry, or even iPhone.
Today this is a non-starter because device development by the handset manufacturers and the wireless providers is difficult. When a new device is created by a handset manufacturer today, the operating system that runs on it has to be optimized with device drivers specific to the hardware on that handset and software specific to that carrier. The manufacturer may have build processes that help speed development, but it's still no walk in the park.
This process takes a long time because while you may have one model of device that can be targeted to multiple carriers, things like OS software that supports transceiver chip sets may have to change in order to accommodate the network technology being used as well as a myriad of other things that make a new device launch on a carrier a big hassle.
So for example, launching the exact same model of Blackberry Bold on Verizon versus an identical model BlackBerry Bold on AT&T could require a lot of customization work. The same goes with any other smartphone.
But let's just say we brought in virtualization, just like we use in today's datacenters. Virtualization is the de rigeur way we solve large numbers of migration and scalability problems in the modern enterprise, but it's a new concept to embedded systems.
With virtualization, we would be introducing Type 1 hypervisors into smartphones. This means that companies like Microsoft, Google, Palm, Nokia and Research in Motion could develop and sell their OSes to the consumer as software products, instead of having it "Preloaded" or baked into a specific device.
As it stands today, I have to buy a Motorola DROID on Verizon with Android. But hypothetically, let's say I want Windows Phone 7 Series on it instead.
Today, that would be difficult, because Microsoft, Motorola and Texas Instruments (which makes the OMAP chipset and platform that the DROID uses) would have to work very closely to develop a Windows Phone ROM image for DROID. It's doable, but it's not easy, and they'd probably create an entirely new smartphone to do it. It also has political implications because Motorola is currently strategically aligning itself with Google and may not even want to talk to Microsoft.
I could either do this myself, or the customer rep at the Verizon store could do this for me, with SD cards sold with the image pre-loaded with it on the "Software" rack at the store. Or do the entire transaction even over their next-generation LTE wireless network or Wi-Fi.
The hypervisor boot screen would say "Hello, I see you have Windows Phone OS Virtual Image for ARM Processors (version 7.0.1) on your SD card. Would you like me to initialize the system and copy the image into flash memory?"
Then bang, 20 seconds later, I have a Windows Phone 7 DROID. Or a Blackberry DROID. Or a Symbian DROID. Or the latest and greatest Android 2.1 software, without having to wait for Verizon and Motorola to develop one, or for the hacker community to build one for me, like the DroidMod or Cyanogen guys are doing today.
Heck, I would have my choice of Android images, from different companies with different features pre-loaded. Maybe my enterprise has a specific image with apps ready to go, with specific security profiles they want me to use. I could switch to a different smartphone OS just as easily as I could switch my ringtone.
And they would all "Just work", just the same way I can copy a Windows 7 or Red Hat virtual machine on VMWare on a HP system to an IBM system in a datacenter with complete transparency. It also means I could export my Virtual Machine to another device on another carrier if I break my phone or want to upgrade to newer hardware.
The above video is a rather technical demonstration on how embedded devices can be built with Intel's Wind River hypervisor and run MULTIPLE operating systems simultaneously. It's geeky, but it proves that this technology exists and could be easily applied today or in the near future.
Primarily Wind River is initially selling this technology to customers who need to run their VxWorks and Linux OS in parallel, such as in automobile embedded computers. But I spoke to them in January of 2010 and they told me that a "very large handset manufacturer" was extremely interested in using the technology to run Android and Windows Mobile side-by-side in the 2012/2013 timeframe.
What does this mean for the Apple lawsuit? Well, it means that if HTC were to adopt hypervisors, all they need to do is sell "blank" phones to carriers. They can get out of the business of customizing and building ROM images and specific phones with specific OSes. It means the consumer can buy whatever OS they want to run on whatever model and form factor of HTC phone they want.
It means Apple can't sue HTC or any other company adopting this technology. It would have to go after Microsoft and Google itself. And it means you would finally have some real choice in what smartphone you want to buy and what you want to run on it.
Will embedded hypervisors on smart phones usher in a new age of customer choice and nullify Apple's lawsuits? Talk Back and Let Me Know.
Updated: Mark Hermeling of Wind River responds on his blog with some interesting comments.