Client Hypervisors: Intelligent Desktop Virtualization too clever for its own good?

In 2011, we asked if  Client Hypervisors will drive will the Next Generation Desktop. Yet, other desktop virtualization industry experts, such as Ron Oglesby, decided the technology was a dead man walking, writing off Type 1 Client  Hypervisors.

Fight? Fight? Fight?

While VMware moved away from client hypervisors, they had to agree that an end user compute device strategy must encompass non-VDI. Their Mirage technology can be considered desktop virtualization, but it is not a client hypervisor. Client hypervisor vendors such as Citrix (who subsumed Virtual Computer’s NxTop) , MokaFive, Parallels, Virtual Bridges and joined by Zirtu. Organisations like  WorldView look to innovate on desktop vitualization through containers rather than full virtualization.

Tablets. Touch Screen capable laptops. Hybrid devices with detachable screens. The netbook might be dead, or they could just be resting. The presence of tablets has undeniably shaken the netbook market but businesses still need powerful, capable laptops.

Bring Your Own Pencil aside – there is still a need to manage “stuff”: still large and small organisations who need to manage the delivery of IT including the end device. The question remains how are devices, and the all important data and applications on them, managed? Hosted and session based desktops have their place – but offline capable device requirements will remain.  Is Intelligent Desktop Virtualization the same as client hypervisors?

What is Intelligent Desktop Virtualization?

Intelligent Desktop Virtualization is more than having the sense to believe that VDI solves every desktop requirement. Fundamentally it is a concept promoted by Intel to maintain central control – yet use the local compute resources. Intel suggest this is achieved through three tenets:

  • Manage centrally with local execution: think desktop virtualization, and there is the immediate thought of VDI – hosted desktops. For the majority of users, such a solution not a problem. But for two issues, the back-end supporting infrastructure can be alien to some organisations or complicated, and/or there is a business requirement for off-line use. 
  • Deliver layered images: fundamentally, separate your user from the OS, ideally the OS from the applications – ideally they are all interchangeable with minimal effort. The reality is this is the key, this the fundamental piece. There are vendors who sit in this separation space – AppSense,Citrix, Liquidware Labs, MokaFive, Unidesk, VMware, Virtual Bridges, World Desk and Zirtu etc. 
  • Use device native management – both physical device management and virtual environments need to be managed for a complete solution. Here, an organisation should be capable of utilising hardware resources independently of the OS to ensure a robust environment.  

These points suggest Client Side Hypervisors can deliver. Yet, device native management is complex for Type 1 hypervisors (where the hypervisor sits beneath the user’s operating system) and not available for type #2 (where the hypervisor is an application within the device’s main operating system). As a Type 1 hypervisor, Virtual Computer’s approach for NxTop had been doing well in getting to grips with integration and delivery to the latest Intel chip features: but in becoming Citrix’s XenClient Enterprise, many customers perceived that early XenClient limited range of devices supported persists. Importantly, the “manage centrally” requires servers and change in delivery mechanisms. While client hypervisors do deliver on a range of use cases, they don’t support all device types. While the scale of back-end infrastructure  is much less than VDI the process of creating images and syncronization needs consideration in very distributed environments.

Ultimately IDV is a journey, with Intel increasingly enhancing their Active Management Technology so that administrators have direct secure management functionality before any of that operating system nonsense begins. Delivery mechanisms need to be able to accomodate these new features – otherwise management environments become a mashup of overlapping consoles  and services.

Too Clever, or not Clever Enough?

Being intelligent about desktop virtualization is key when looking at transforming your desktop environment. Even pure hosted or session desktops still require a device for the users to interact with, for many organisations that device needs to be managed in some way. We could consider from an Intel point of view, hosted desktops don’t sell more expensive end-devices: it could be considered a parry to the thrust of hosted desktops. Yet, consumer/prosumer drive is for mobility in a device form factor – laptops, tablets, hybrids. However there are still many sizeable, or critical environments where a user owned portable device doesn’t make sense: be it for class room delivery, call taking, administration.

Do Client Hypervisors address the three tenets? Manage centrally with local execution: yes. Deliver layered images: yes. Use device native management: ish.

There are a number of solutions that allow you to manage enviornments centrally with local execution: with such as Microsoft’s InTune or Bravura’s OptiTune being offered as cloud services. In terms of layering, file syncronisation services are moving beyond Dropbox to offer enterprises layered access to data with consumer flexibility direct to an organisations data – Citrix’s ShareFile,  or EMC Syncplicity for example. Device native management is the more complex.

We’ve said before that the next generation desktop would need to:

  • Be more responsive to business change.
  • Be transferable between devices.
  • Be recoverable and supportable, regardless of location
  • Allow layering of services

Client hypervisors have an important place for a number of use cases. While legacy environment support is small, it is still important: Windows XP ‘s support might be ending, but it will likely still need to run. Beyond that there are development and testing environments; class room environments; environments where there are a high turnover of workspaces or requirement to be able return to a steady state on the end device.  An important goal is to make the end environment simpler, not more complex. To the next generation desktop deliverables there should also be “make it simple” explicitly in the mix somewhere. With some intelligence, client hypervisors are capable of reducing complexity – but it is possible to get it wrong.

Intelligent use of desktop virtualization is indeed key. There are many new and interesting technologies, but it isn’t about the new and interesting technolgies – its about delivering the data and the applications. If the end environment you are delivering is as complex, or more complicated than what you have now: crumple up the page, start again.