Smartphone Virtualization: Useful Or Useless?
The virtualization industry has been on fire with news of VMware's latest experimental venture: Smartphone Virtualization. The technology has been targeted at the mobile professional who currently carries two phones, one for personal and one for business use. VMware touts that with a real hypervisor on board, our prototype telecommuter of tomorrow will be able to switch seamlessly between two different installed operating systems on the same phone. With Gigahertz processor cell phones on the market and storage into the 100GB mark, it seems reasonable to virtualize phone technology. Getting deeper into the logistics, however, it starts to sound like more trouble than it's worth.
First, mobile computing is an infantile industry. Although there are some shining stars like Apple's iPhone and the Google Android OS, in general, mobile computing continues to be a cumbersome and difficult to use technology. Interface points are small and cumbersome, battery usage is ridiculous on application enabled devices and internet speeds are just barely into the usable range on the best of providers. No matter who you are, getting all the features you like on a single device seems to be impossible. For example, I love the application support and interface fluidity of the iPhone, but the touch keyboard is absolutely impossible to use easily for someone with reasonably large fingers. Conversely, the Blackberry 8800 series phones have a tremendous keyboard but the interfaces are cumbersome and unpleasant even if there are a fair number of apps available. Add to this the fact that considerable configuration needs to be done before the phone fills all the requisite roles, and you have a real headache.
Enter VMware with virtualization for your phone. Now, instead of one largely unmanageable and cumbersome phone operating system, you have two. How will you switch between the business and personal phones? How long will it take in seconds before your business phone is available after pushing the button, and what happens when a call comes in on the other phone while you are occupied? What will VMware do with the inactive phone image, and what will the ramifications of it's suspension be? If the image is not suspended completely when you hit the button, then it will continue to use CPU resources, sapping vitality and power from the active image. Second, what about the logistical aspects of this new technology? Backing up cell phones that are virtualized may be the same as saving a .vmdk file, but what application is going to communicate with all these different devices to pull the disk files? And what about general resource usage? Most of the phones I've used don't run all that fast even with exclusive hardware access. What happens when the inactive image is chugging away in the background? If we shut the inactive down completely then it can't take calls and we'll have to power it back up with a substantial delay before we can use it. A short term thinker might say, "Don't worry, cellular CPUs will be faster in a year or two and better able to handle the dual loads."
This type of thinking is faulty for two reasons. First, as time passes, our mobile applications are becoming more demanding. This trend will continue to be positively correlated with the progress of technology. More powerful and demanding apps will require more powerful phones, partially or wholly neutralizing increases in CPU speed. Second, CPU speed is correlated with power. In the laptop market we see this consistently as more powerful machines use larger power supplies and more significant heat sinking. This is not a workable model for cell phones which are already barely battery efficient. Although there have been promises of new battery technologies on the horizon, the last commercially viable battery improvement to hit the market was Li-Ion in the 1970's!
Finally, how will the question of phone ownership and service delivery be resolved? Regardless of what phone platform, both the business and personal phones will be tied to the same service provider dramatically reducing consumer options. Who will pay for the shared device, the office? The consumer? In what tedious and complicated way will we settle the partial usage of the device and the airtime? Does anyone else think that this adds even more complexity to an unneccessarily complicated situation?
Jasmine McTigue is the IT manager for Carwild Corp. She is responsible for IT infrastructure and has worked on numerous customer projects as well as ongoing network management and support throughout her 10-plus-year career.