Misc, Virtualization, writing
Leave a comment

Future of OS virtualization

In this post I will write about my thoughts on how I wish or perceive the OS virtualization technology to be in future. Over here, I will talk about how the hypervisor layer might look like in the future and as a result, how a few more things around the hypervisor will change. I will also talk about a few benefits of this new evolution.

 All the information in this post is looking at the future and probably how I envision this technology. There may be companies or people already working along the same lines. I DO NOT have knowledge of any such “confidential” information. If I were, I will not discuss that here.

 The Hypervisor

If you look at the evolution of virtualization there were multiple phases in it. It all started with mainframes and moved on to hosted virtualization or type-2 hypervisor products like Virtual PC, etc. Next biggest leap in this technology area was the development of type-1 hypervisors or bare metal hypervisors such as Microsoft Hyper-v or VMWare ESX. Today, we have embedded hypervisors – as tiny as 32MB – running out of a SD card.

 In future, I believe that this hypervisor would become much thinner and eventually become a part of platform firmware such as EFI or uEFI. This will enable multi-tenant architecture support right at the hardware level. At a high level, it may look something like this.

Figure 1: Hypervisor moving in to the platform firmware

Figure 1: Hypervisor moving in to the platform firmware

Now, if you look at the above figure, the “Hypervisor Boot Services” piece could be a vendor-neutral layer that is standards driven and developed jointly by companies like Microsoft, VMware, Red Hat, Novell, Intel, AMD, etc. This layer will provide the required base services – such as IO scheduling, memory management — of a hypervisor. OEMs and ISVs will provide the extensions that will enable management of bare metal system, hypervisor and guest OS running on top of the hypervisor. These extensions can also be used to implement various differentiating features from each of the current hypervisor vendors, OS vendors and hardware vendors. 

There will be various benefits of this implementation. A few of them will be

·         Support for any guest OS – A firmware implementation of the hypervisor gives way for running any guest OS as long the drivers required to access the hardware / hypervisor services are available

·         VM packaging and deployment – All management interfaces will be standards based hence, there will be a consistent method to package and deploy virtual machines. Open Virtualization Format (OVF) can easily be extended and used in this new model.

·         Hypervisor for consumer PCs – The same firmware implementation will be used on both enterprise and consumer platforms. Hence, making it possible for consumers to run multiple operating systems of their choice without any need to use hosted virtualization products.

 

Virtualization-aware data centers

The biggest benefit of this new hypervisor implementation will be the evolution of virtualization-aware data centers. Along with other boot & runtime services implemented within the platform firmware, every system — within the data center — hosting virtual machines will be capable of providing systems management data such as resource utilization, power consumption, etc.

 

This data can, then, be fed in to various other systems like systems management software, network infrastructure, etc to improve the overall QoS and uptime of virtual machines and host systems. These virtualized data centers will be able to use the built-in capabilities of the hypervisor and OEM / ISV extensions to deploy cloud services on demand.

 

Who will win?

Okay, I have talked so far about moving the hypervisor layer in to the platform firmware and making it standards driven. Does that mean all of the present virtualization vendors will be jobless?

Not necessarily. These vendors get to concentrate more on the value additions they can provide and of course, the management aspects of a virtualized data center.

 

These vendors will have to concentrate on interfaces that will

·         Enable consistent method to manage both virtual and physical worlds. Customer will look for ways to manage OS and apps from a single console

·         Enable seamless migration of virtual machines from one host to another

·         Create efficient cloud services for the virtualized datacenter

·         And many more….

 

The value additions and the differentiators make a few of the existing companies successful in building a story for the dynamic data center. This decides whether these companies win or lose.

 

PS: If you are wondering why I started talking about all this now, read this http://www.facebook.com/topic.php?uid=62844697987&topic=9278. This is my entry in the virtualization category. There may be technical inaccuracies or unclear implementation details. This is just my vision and not a design document. Please leave your comments here on what you feel about this post and the content. Your thoughts may be useful for me in refining the whole article

Filed under: Misc, Virtualization, writing

by

Ravikanth is a principal engineer and the lead architect for Microsoft and VMware virtualized and hybrid cloud solutions within the Infrastructure Solutions Group at Dell EMC. He is a multi-year recipient of Microsoft Most Valuable Professional (MVP) award in Windows PowerShell (CDM) and Microsoft Azure. Ravikanth is the author of Windows PowerShell Desired State Configuration Revealed (Apress) and leads Bangalore PowerShell and Bangalore IT Pro user groups. He can be seen speaking regularly at local user group events and conferences in India and abroad.