Virtualization – A Feature of the OS, or a New Platform?

Microsoft and Red Hat have just announced that they have completed the certifications of the cross OS hypervisor agreements that the companies originally announced back in February of 2009. This means that Microsoft now certifies Red Hat Linux guests on Hyper-V and Red Hat certifies certain Windows guests on KVM. Red Hat has an excellent article on its web site that details which version of which products work with which and which provides an excellent FAQ.

This tactical level of cooperation between the two operating systems competitors/collaborators highlights an extremely important issue in the system software industry. That issue will literally determine the future of virtualization. This issue is explored below:

Is Virtualization Just a Feature of the OS?

Clearly Microsoft and Red Hat would like it to turn out this way. In the eyes of both operating system vendors they have made the investment in hardware drivers and compatibility, have a features that allocate and schedule CPU, Memory, Network and Disk resources, have a rich set of applications services and API’s and have substantial installed bases worldwide. The view of Microsoft and Red Hat can be neatly summed up in the image to this effect on the Red Hat web site.

MS_Red_Hat_Hypervisor_Interoperability

Notice that there is no bare metal hypervisor underneath the OS in this picture. The view on the part of both Microsoft and Red Hat is that such an extra layer is an unnecessary source of complexity and expense, and that virtualization is rightly just a feature of a modern operating system. In order to prevail with this point of view, Microsoft and Red Hat must accomplish the following:

  1. Effectively leverage their installed bases and the administrator skill sets in those installed bases and deliver “good enough” virtualization to these constituents without having them go through and their employers pay for another layer of training and certification just to install virtualization in their environments.
  2. Address the critical gaps in functionality that exist vis-a-vis VMware vSphere. These include memory page sharing (which directly drives guest density), and the vStorage, vNetwork and vmSafe interfaces which third party vendors use in VMware to add significant value to the platform.
  3. Court the ecosystem of third party software vendors so that they fill in the remaining gaps which exist between the Microsoft and Red Hat offerings and “Virtualization Suite” from VMware.
  4. Cooperate on not just the hypervisor and guest front but also the management front. Put in place common interfaces and in some cases common tools that allow these two environments to be managed in a similar and leveraged manner.
  5. Find a way to leverage the ecosystem of VMware VAR and SI partners. These implementation and services partners really do not care what the price of the software is, since in large deals their margins on the software are compressed to the point of little impact upon their businesses. What they do care about is the services profile of these solutions. VMware delivers a high ROI to the customer despite the fact that VMware projects are expensive from a services standpoint – because they are staffed with expensive VMware certified engineers. VMware and Red Hat need to find a way for this partner community to make as much or more money with their virtualization projects as the partners do with VMware projects – while requiring less skilled engineers, who bill out a lower rates.

Is Virtualization a new System Software Platform?

Certainly VMware would like to have you believe that VMware is at the minimum a new “layer” that gets inserted between the OS and the hardware, through which a variety of new services get delivered to these containers in a manner independent of the OS and the application running in the container. VMware also intends to take this further as was discussed in this article about VMware “No OS Application Platform Strategy”. The end game here is to glue applications run time platforms (like the recently acquired SpringSource) directly into a VMware Guest, removing the need for the underlying OS.

app_architecture_no_os

In order to succeed in establishing itself as a new systems software platform, VMware must accomplish the following:

  1. Clearly position, package and price the virtualization platform differently from the management stack for the platform. VMware is furiously building out a stack of management extensions to its hypervisor. These include DRS, HA, FT, SRM, Lifecycle Manager, Lab Manager, AppSpeed, as well as the forthcoming Capacity IQ and Config Control products. These management extensions add great value to the VMware platform (and provide VMware with things to sell that Microsoft and Red Hat have no current answer for). But the focus on the management extensions risks causing VMware to takes its eye off of the most important ball of all – which is to expand the market share of the platform in all segments (small, medium, and large – across industries) of the market.
  2. Show the ISV ecosystem where the opportunities are to add value. Right now, the third party ISV ecosystem looks at VMware and sees a management stack strategy that looks like it is designed to drive them out of the virtualization market. The truth (whether VMware realizes it our not) is that it will be impossible for VMware to build a management stack that will meet the needs of every market segment that adopts (or could adopt) its virtualization platform. If VMware can find a way to be clear with the third party vendors where VMware will not compete then VMware will have the best of all worlds – its own high value management stack, AND a vibrant third party ecosystem.
  3. Show the implementation partner ecosystem how to make money in mid-market accounts so as to mitigate against the market share losses that will otherwise occur in these segments due to the natural affinity on the part of Windows shops for a Windows oriented virtualization solution.

Extending Virtualization in the Enterprise

As was laid out in this article about the Management Stack Strategies for Microsoft and VMware, enterprises are recommended to take the following steps to maximize their value and choices from virtualization initiatives:

  1. Decide if you are going to virtualize tier one applications or not in the next 12 months. If you are then VMware has in insurmountable lead within this timeframe as a virtualization platform for those applications and you should proceed with your likely existing plans to virtualize these applications on VMware vSphere.
  2. Decide if you are going end up being a multi-hypervisor company or not. This involves trading off the simplicity of having one virtualization platform against the value of having negotiating leverage with multiple vendors.
  3. If you are going to end up with multiple virtualization platforms, then start figuring out what those platforms are going to be (Microsoft, Red Hat or Citrix) and which of the low hanging fruit environments (test and dev) and applications (tactically/department Windows based) that you put on VMware two years ago are now candidates to be moved “down” to your second and presumably cheaper virtualization platforms.
  4. If you are going to end up with multiple hypervisor platforms, decide where to draw the line above which you will pursue cross-platform management tools. This is essential so that you do not end up with two stacks of competing and non-integrated management tools for your virtualization environment.

In summary, this is a once in a 20 year life or death struggle for dominant position in the systems software industry. Both Microsoft and VMware come to the party with incredible strengths. This is going to take five years to play out, and should be very challenging and entertaining in the process.

Bernd Harzog (318 Posts)

Bernd Harzog is the Analyst at The Virtualization Practice for Performance and Capacity Management and IT as a Service (Private Cloud). Bernd is also the CEO and founder of APM Experts a company that provides strategic marketing services to vendors in the virtualization performance management, and application performance management markets. Prior to these two companies, Bernd was the CEO of RTO Software, the VP Products at Netuitive, a General Manager at Xcellenet, and Research Director for Systems Software at Gartner Group. Bernd has an MBA in Marketing from the University of Chicago.

Connect with Bernd Harzog:

Tags: , ,

Leave a Reply

Your email address will not be published. Required fields are marked *

Please Share

Featured Solutions