With today's mobile app stores & Google Voice, the mobile phone hypervisor is dead

When I first started using a mobile phone in the late 1990s, I still had a "work" phone at my desk and a landline phone at my home. Over the years I began to use my mobile phone more and more, and eventually it became the only phone I had (for both work and personal use).

When I first started using a mobile phone in the late 1990s, I still had a "work" phone at my desk and a landline phone at my home. Over the years I began to use my mobile phone more and more, and eventually it became the only phone I had (for both work and personal use).

But in 2007 I realized that having a single phone for work and personal was not good. The main problem was that it just had a single phone number, so the girl I met at the bar that night had the same number as a potential new client at work. I realized that I needed to separate my phone lives, and this being 2007, the only choice I had was to get a second phone and to dedicate one to personal use and one to work use.

Of course at first i just got a basic personal phone, but I found that I missed being able to have maps and Google with at all times, so I'd end up carrying both my work and personal phones with me anyway. After a few months of this I bit the bullet and upgraded my personal phone to a smart phone which meant that I only needed to carry a single phone with me during non-working hours, but I still had two phones with me at work because I didn't want to miss any personal calls.

You can imagine that when VMware bought mobile phone hypervisor maker Trango in November 2008, I was a happy guy. I was also happy a few months later when Citrix invested in a similar company called OK Labs.

I remember being excited at the idea of running two phone OSes side-by-side. "At last!" I thought, "One phone, two identities!"

Of course the initial excitement of 2008 faded when the vendors reminded us that any practical phone-based hypervisor was "years away." And probably within a few months of seeing the demo, I completely forgot about the concept of the mobile hypervisor and I moved on.

Fast forward to October 2010...

A few weeks ago, my company (TechTarget) changed their policy around mobile phones and now allows full Exchange ActiveSync with Android- and iOS-based devices. (Previously they were Blackberry-only.) Up until then, I had been using a Blackberry for work and an Android for personal use. (I used to have an iPhone, but it was pretty worthless in San Francisco, dropping about 1/3 of my calls. I switched to a Motorola Droid on Verizon.)

Once we started supporting Android, I hooked the Android phone into our corporate email. It was a pretty quick process and gave me all the usual Blackberry-like features (contact & calendar sync, push email, etc.). And since Exchange ActiveSync also allows employers to enforce device security, I had to accept their terms when linking my phone to my work email account for settings like enforcing device lock timeouts and allowing them to remotely wipe the device if I lost it.

So now I have my full work email experience on my personal phone. This is nothing new. What's new for me is that I was able to transfer my full work telephone experience to this new device too.

For the past six months or so, I've been using a Google Voice number as my primary work telephone number. Google Voice essentially virtualizes your phone number--people who want to call you use your Google Voice phone number which you configure to ring on one or more physical phones.

In this case my new Android phone can do double-duty: Personal calls can come in directly via its native phone number, and work calls can be routed to it via Google Voice. (An added benefit of Google Voice is that you can configure "do not disturb" hours, so that business callers using the Google Voice number only ring through from Mon-Fri, 8:00am-6:00pm. Outside of those hours, they go straight to voicemail.

The Google Voice integration is simple and seamless for people who call me. The problem so far has been that when I call people using my mobile phone, the caller ID always shows as the phone's native number. This is a problem for two reasons:

  1. If the person I'm calling only has my work number in their phone, then me calling them from my personal phone number means that I show up as an unrecognized number.
  2. Since I'm calling from my native phone number and not my work number, my native number is exposed to the person I'm calling. If they save that number for me, then they could call me back directly and bother me when I'm out with friends, sleeping, etc.

The solution to this is to install the Google Voice Android app. It has the ability to enhance the Android's native dialer so that it, for example, always uses the Google Voice phone number as the caller ID number instead of the phone's native number. (You can also configure options for whether you should dial via VoIP or normal cell phone.) In my case I configured the dialer to pop up a prompt to ask me which number I want to call from, essentially allowing me to select which caller ID I show when I call someone.

A great solution with no client hypervisor in sight!

So right now it would appear that I have the ultimate solution. I only carry a single device (my Android, which my company pays for). I have access to both my work email (via Exchange ActiveSync) and my personal email (gmail, which the android is able to keep separate from my work stuff). I have my calendar, my contacts, push notifications, and full device backup to my Google account. And of course, I have the ability to both make and receive work and personal calls from the single device while controlling which calls go where and when. And most importantly, since it's "my" device, I can also install whatever other apps I want on it. (This includes real work apps like Concur Expense & Travel and the Citrix Receiver, as well as time-wasting fun apps.)

Even more importantly, since I'm connecting to my work email via Exchange ActiveSync, my employer can enforce the security of my device remotely.

So I ask: Why would I ever need a client hypervisor on a phone?

The initial use case that occurs to me is for situations where I want to carry a single device while running a different OS for personal use than the OS my employer wants to use. But I wonder... does my employer really care what OS I use on my device? My guess is "no." I assume that as long as my device is compliant with what they need to do (in this case the ability to be managed as an Exchange ActiveSync device which allows for remote security enforcement), then I would assume that my employer couldn't really care less about which OS (or even which make or model) my mobile phone is.

Running many different apps side-by-side on a mobile phone is simple compared to running multiple apps on a Windows desktop. The mobile apps (whether they're for iOS, Blackberry, Windows Mobile, Android, Symbian, WebOS) all share the trait of "playing nice" together. And they're all now available via mobile phone app stores that handle the complexities of provisioning, updating, data storage, and deprovisioning. So IT departments don't really have to worry about app management on mobile devices like they do on Windows desktops.

Combine that with things like the Citrix Receiver and the concept behind Citrix's Project Golden Gate, and the actual device becomes even less important. It seems that the only real sticking point for folks is how to handle multiple phone identities on the same device, and services like Google Voice take that challenge out of the mix too. (And for the record, there aren't even any mainstream phones that could take multiple SIM cards to allow a mobile hypervisor to mix & match identities.)

So while Citrix and VMware work on their mobile hypervisors in the labs, it seems that the world has moved on. I personally am enjoying everything I thought I needed a mobile hypervisor for today without one.


Join the conversation


Send me notifications when other members comment.

Please create a username to comment.

Running multiple OS on a phone concurrently does not work anyway because of the lack of resources. A phone is fine-tuned to provide exactly the right amount of power and battery capacity - introducing a hypervisor would break that delicate balance.

I agree with Brian in that I cannot see a use case for a mobile hypervisor.

An additional thought: would Apple and all the others give up some of the control they have over their platform to the hypervisor maker? I do not think so.


I've seen Good fo Enterprise in action http://good.com/ (I'm sure there others, I'm have no affiliation with Good). If your device doesn't meet corporate policy then install this app which does meet policy - corporate data is then seperate from personal and under different security policies.

I didn't see the benefit of a mobile hypervisor in 2008 and it makes less sense now.


I agree with your assessment, but I applaud VMware and Citrix for placing some bets on mobile hypervisor technology.  It could very well be dead, but it's possible that with a major mobile phone security debacle or two it could be back to life in a hurry.  Even if that were to happen, I see it as an OEM play to the handset manufacturers or mobile OS providers versus a traditional virtualization product model.

I recently implemented a poor man’s version of Brian’s model by porting my home landline to a service called 3jam.  This is similar to Google Voice, but unlike Google you can port an existing number to it (though it’s not free).

I wrote up the process I went through to do this on my personal blog:


If you don’t care about porting your number, the Brian’s Google Voice approach is definitely the better way to go, particularly given that it is free.  Even for iPhone users, the Google Voice native apps are starting to resurface again.

With the 3jam approach I took, you get all of the same inbound call handling capabilities as Google Voice, but there isn’t as much support for mobile outbound calling with the correct caller ID.  In my case, I hacked it by changing my Skype outbound caller ID to my virtual number.  So, I can make an outbound call from Skype either on my iPhone or computer, and my old landline number shows up on the caller ID.



I would argue that the hypervisor market for mobile devices is actually the most widely deployed devices compared to desktops/laptops and servers. OK Labs has there OKL4 microvisor embedded on over 750 million devices. That topples client hypervisors and I am not sure on the number of server hypervisors are out there...

Remember the rant on how virtualizing a server is not the same as virtualizing the desktop? They are completely different technologies with completely different benefits, this holds true for the mobile devices as well.

Who cares about running multiple OSs on a phone, the key is about running isolated secure applications side by side and virtualizing the device at the driver level.

This post is actually good timing because OK Labs just recently announced SecureIT, a solution that enables the commodity of COTS hardware being utilized for secure wireless communication.



Oh that's really interesting. You know, I only thought about the mobile hypervisor in terms of multiple simultaneous VMs since that's what VMware showed on stage... but while that demo was cool, it was also maybe misleading since people (well, me) might think about that was the only use case.

But if it's really about hardware abstraction and patch management, and if OK Labs is already on 750m devices, then I guess the mobile hypervisor isn't "dead" after all. Maybe I should write that the "use case of multiple mobile OSes is dead."



I would agree if this were purely a technical issue.  

As it stands, even with the recent tax code changes there are still serious tax implications companies have to consider for employer provided cell phones.  UCLA was just fined a couple of years ago to the tune of ~$120,000 because they failed to properly account for personal use of corporate owned cell phones.  I know several colleagues that still carry two cell phones specifically to avoid any potential conflict with the IRS.

By having the ability to tie two contracts to a single device a company, or person, can indemnify themselves against being taxed for personal use of corporate assets.



Hopefully you didn't think I was blasting your article, I really wasn't.

This space is good to watch out for, especially since Intel bought McAfee and wants to make their chips superior to ARM and get in this area, this might lead to new innovations.


I’ve carried multiple devices long enough that I’ll deal with the occasional “oh, I’m on vacation, at the hospital, etc. rather than maintain the charging stand and utility belt.  

I don’t ever expect to have a  Mr. Rogers type job (if you are under 25, you probably won’t understand this) where I can take off the wild jacket and shoes and put on the work attire and then change back before I leave my workplace.  Maybe I’m a bit unique in the grand scheme of things though.   I think I’d find it exhausting trying to be two totally different people during the day and night.  

However, I think Brian is just stirring the pot because that’s his job.  Brian, you’re a “Pot Stirrer”.  ïŠ   I do believe there’s a market for a phone hypervisor just as I believe there will be disruptive technologies beyond the blackberry, ipad, iphone, android, etc. that’ll change the end user computing landscape and once again, bring the consumer devices into the office, off the corporate book of assets but into the realm of use and productivity for the employee and employer alike.  These changes will bring challenges and struggles for control by corporate IT and the owning consumer.  A hypervisor of sorts seems like a logical solution but I don’t think it’ll be used for the same benefits as we see on the servers or desktops.  The uses cases are being developed as obstacles are discovered and security, compliance and productivity concerns are identified.  



I guess I still don't "get" how a mobile hypervisor is needed. For example, in one of Simon's posts, he talks about the value of app developers targeting the OS, but the hypervisor allowing that app (via a VM with the proper OS) to run on any device. (So it's the classic "any device, one app" advantage for the developer.) But why do we need a hypervisor for this? Isn't that killing an ant with a sledgehammer? Wouldn't this be where HTML5 or AIR or other RIA platforms would allow developers to write once / deploy everywhere without needing a hypervisor?

And then the example of how a malicious app could steal your contacts and spam them without your knowledge, that happened because users installed an app where they granted access to that information, but the app itself was malicious and did things with the info that the users weren't expecting. But again in this case, I don't understand how a hypervisor would prevent that? If I download an app that does something cool but that needs access to my contacts and email, then I'm going to give it access. And if it does something bad, then there's nothing the hypervisor can do about it, right?

Honestly the best argument I've heard so far for the multiple VMs is for the separation of work/personal billing.. although honestly by the time these things come out it might be all IP data anyway, which would be easy to sort out from the network end and wouldn't require multiple VMs.


Agreed, both HTML 5 and Air will go along way to reduce the platform diversity issues, but it will only reduce not eliminate the problems.

But neither address the security issues of user installed rougue apps  accessing enterprise data. When the goal is to place multiple personas on a single device you have to comprehensive isolation between the two. Given the almost ubiquity of OK Labs microvisor and ARM Cortex processors we have an environment that is predisposed towards the use of hypervisors as a baseline configuration.