One topic that gets discussed quite often is Microsoft Hyper-V vs VMware vSphere, and a quick Google search for comparisons will return at least several hundred thousand hits. There seems to be a large number of posts and articles trying to make a case that one is better than the other by listing and comparing features of the hypervisors themselves one by one. The purpose of this post is not to claim that one platform is better than the other. Is that the best way to really compare the different virtualization technologies as a whole, or should we take a step back and really look at differences in approach for the virtual infrastructure and/or virtual ecosystems?

Microsoft:

In my opinion, Microsoft is defining and building their virtualization infrastructure as an extension or expansion of their current ecosystem, with System Center at the center of their universe. If you look at the System Center 2012 product page on Microsoft’s website, System Center product details are broken down into two different areas:

  1. Cloud and Datacenter Management
  2. Client Management & Security

Is this really a big surprise? Absolutely not, since it clearly makes more sense to build on what you already have in place than to reinvent the wheel. The majority of virtual machines that are running on the Hyper-V platform are running Windows, and System Center already has a solid foundation of features and capabilities for managing Windows environments. These features include:

  1. Application Delivery
  2. Mobile Device Management
  3. Virtual Desktop Management
  4. Endpoint Protection
  5. Compliance and Setting Management
  6. Software Update Management
  7. Power Management
  8. Operating System Deployment
  9. Client Health and Monitoring
  10. Asset Intelligence
  11. Inventory

VMware:

In my opinion, VMware is looking to create a completely isolated and separated ecosystem that consists of a collection of appliances with different capabilities working independently and making up the features within the infrastructure, including:

  1. vSphere
  2. vCloud Director
  3. vCloud Connector
  4. vCloud Network and Security
  5. vCenter Site Recovery Manager
  6. vCenter Operations Manager Suite
  7. vFabric Application Director
  8. vCloud Automation Center

Conclusions:

One of the main differences that I see in the two approaches is that Microsoft wants virtualization, cloud, and datacenter management to be an extension of the infrastructure, whereas VMware would like the vCloud Suite to be the complete infrastructure. This starts with VMware developing vCloud as an Infrastructure-as-a-Service to fulfill their promise of the software-defined datacenter.

These are two completely different approaches that seem to define which segment of the industry each company has their eyes on. Microsoft appears to be most interested in the small to midrange enterprises, and VMware seems to be focused on the larger enterprises. From my experience throughout my career, in smaller environments administrators are expected to play a greater role in maintaining a larger scope of different technologies in their environments, or in other words, to be “do it all admins”. The larger the organization, the more compartmentalized the staff and technologies become, relying on Subject Matter Experts (SME) for all the various technologies, including Networks, Database, AS400, Midrange, Messaging and so on. With Microsoft focusing on extending its infrastructure, it appears to be catering more to the smaller and less compartmentalized organizations. I do not mean for this statement to be taken as a rule, for both Microsoft and VMware have been working hard to make sure their products are able to support and to scale for both smaller and larger organizations, and I am sure there will be plenty of smaller organizations that will be running vSphere, as well as larger companies and hosting providers that will be running Hyper-V, as well as others.

Throughout the years, I have seen plenty of heterogeneous environments with a collection of Linux, UNIX, Windows, and AS400 technologies, and I do not see that changing with virtualization. In fact, I picture a future with multiple hypervisors working together in harmony within clouds. More and more products are being developed to be able to support multiple hypervisors, and I believe this trend will continue and become more prevalent in the near future. I do not think the hypervisor should define the cloud, but rather the hypervisors should enhance the cloud itself.

Picture having the ability to move virtual machines and applications from hypervisor to hypervisor from a central location without ever skipping a beat. The pieces of technology that would be needed are, for the most part, in place, and we are just missing the integration engine to communicate effectively with the different platforms. Both Microsoft and VMware have great tools in place to migrate virtual machines into the respective platforms. We are just missing the ability to manage and create a workflow to migrate at will between platforms using the tools that are already in place. I do not foresee Microsoft or VMware adding this capability into their respective technologies; I would really expect something like this to come from a 3rd party, but I am very much looking forward to the time when it comes.

Software servers, software networks, and software storage all move us along the track to a completely virtual datacenter and as asked early on, how many datacenters do you know that are not heterogeneous in one way or another? This will be the way of the future.

3 replies on “Microsoft Hyper-V vs VMware vSphere”

  1. “Picture having the ability to move virtual machines and applications from hypervisor to hypervisor from a central location without ever skipping a beat.”. I like that concept – and fundamentally it (should) underpin the reason to moving to the cloud.

    Yet, I think the reality (especially for the SMB market) is to rationalise not expand out per-se. I agree the “wider cloud bursting capability” will be of use and “of use” to a market that will pay money? Economy of scale comes with standardisation.

    I agree we’re moving to a virtual data centre, and I agree the ideal would be to support a multiplicity of environments. Still, will not the practicalities of service provision be that offer the “best” and “cheapest” service on a single platform?

  2. There are two ideas that conflict with some of that may result between Microsoft Hyper-V and VMware vSphere. 1) Enterprise customers are starting to become monolithic by need not desire. The idea of having only one OS for the desktop and only one OS for mid-range servers is real, and many firms are realizing they cannot avoid jack of all trades SAs or even a large number of SMEs, so the reality is, support many OSes, which only resulted because many vendors only did on OS, is dying a slow, very slow death. Enterprise wants to reduce TCO and drive up ROI? That is done must effectively, and efficiently, by going more monolithic. 2) VMware has a strategic issue, that is coming, if years away, that is that as OSes become more hypervisorized in concept, the raise or increased adoptance of virtual enivornments (VEs), or what we call often call generically virtual containers will increase in popular acceptance. This significantly will reduce the impact of running OS isolation based virtualization such as vSphere, versus application isolation based virtualization such as LXC (for Linux) or Server-AppV (for Windows, well, really the next significant generation of Server-AppV to be fair). This effectively addresses the small and mid-sized firms as having virtualization, without extensive replication of OS instances, key to TCO reduction, for configuration management, patching, compliance, etc., etc. Never mind reduce provision costs, as IaaS gives way to PaaS, and SaaS moves to AaaS based. Enterprise class entities are already tackling PaaS and realizing that OS isolation is expensive now that virtualization is a defacto reality.

    It really comes down to TCO, few OS platforms in frequency as well as type, that offer expected virtualization as demanded, will dominate long term, and VMware just can resolve the challenge of this trend, since they own nor control any OS that is application space based, such as Windows or Linux. Microsoft thinks it can own this trend over time, everything they are doing is geared to this, even if Microsoft will not say so, it is at the heart of every step they have taken for the last 6 or so years. The problem is, the Windows OS is showing its age, but that will change post Windows 8, when Microsoft will finally revamp its server OS strategy, having burned out its eco-system the idea non-core stylized server OS, that supports true VEs, while enhancing desktop oriented management.

Comments are closed.