Who Runs What Hypervisor?

Have you taken any time answering this question? Who runs what hypervisor? Is it just me, or do there seem to be a lot of articles and posts about OpenStack recently, so many that one almost gets the feeling that everything is running on OpenStack? It looks like there’s a push to help keep OpenStack on the path to becoming more mainstream, and the new partnership with Red Hat might just be the ticket. For now, OpenStack is still going through its adolescence, but it has great potential to go out and really make a difference in this world. Until then, have you ever stopped to consider which underlying hypervisors are supporting the clouds we all know and love?

Today we still need to consider those underlying hypervisors when moving workloads between the clouds. The software-defined data center, with a few exceptions, often requires changes to virtual machines depending on the hypervisor in use as we move data from one cloud to another. Here are some popular clouds and what hypervisors they run:

Google Compute Engine (GCE) is Google’s Infrastructure as a Service (IaaS) platform that uses KVM (kernel-based VM) as its hypervisor and only supports guests running on Linux distributions of Debian or CentOS.

Amazon Elastic Compute Cloud (EC2) is the central part of Amazon.com’s cloud computing platform, Amazon Web Services (AWS). (On a side note, in November 2010, Amazon switched its own retail website to AWS.) EC2 uses a custom Xen implementation as its hypervisor. It supports Microsoft Windows, Linux, and FreeBSD running as guest OSs. The hypervisor runs on approximately half a million Linux server instances.

Microsoft Azure is Microsoft’s cloud computing Platform as a Service (PaaS) and IaaS that supports Microsoft Windows and Linux guest operating systems. Azure offers several different services as it attempts to achieve the success and acceptance of Hyper-V. The Windows Azure operating system runs a fabric layer that manages the resources running on Windows Server 2008 and a customized version of Hyper-V.

VMware vCloud Hybrid Service (vCHS) and VMware vCloud (implemented by third parties) are IaaS offerings based on VMware vCloud that run on the VMware ESXi hypervisor. These services are meant to serve as an extension of companies’ private clouds that run on the VMware platform. This allows customers to migrate work on demand from internal to public while remaining on the same base hypervisor. ESXi supports Linux, Windows, FreeBSD, and other guest OSs.

Rackspace Cloud Servers offer a cloud infrastructure service that allows users to deploy from one to hundreds of servers instantly to create advanced high-availability architecture. These Cloud Servers also run on the Xen hypervisor for Linux host instances and Linux guest instances and on the Citrix XenServer hypervisor for Windows and Linux guest instances. In 2010, Rackspace contributed the source code for its Cloud Files product to the OpenStack project under the Apache License, to become the OpenStack Object Storage component. In April 2012, Rackspace announced it would implement OpenStack Compute as the underlying technology for its Cloud Servers product. The change will include a new control panel as well as add-on cloud services offering databases, server monitoring, block storage, and virtual networking.

The point that I want to make is that although there is a lot of news and hype around OpenStack, this platform still has a way to go to catch up to the functionality of the other cloud providers. It is not quite ready yet, though many are interested in it and developing for it already. It has a chance to be the way of the future. But for now, Xen would seem to be champion due to the sheer numbers of physical hosts—approximately 500,000—at Amazon alone. This does not mean you can’t go wrong with Microsoft or VMware. In fact, in my opinion, each cloud service has different strengths and weaknesses, and different companies have different needs. I am not sure the platform offerings will be one-size-fits-all.

If you worked in the industry from around the mid 1990s forward, you might remember the IT family feud between the Windows team and the Linux/Unix team. The competitiveness between the teams led to a lot of “my servers are better then your servers” back-and-forth between teams.  Although this may have been somewhat fun in after-hour events, I am sure that in many cases the best real solution for the customer lost the political battle because teams stuck with their political party (Windows vs. Linux) regardless of what was the best solution overall. What I would really like to see for us in the twenty-first century is to finally leave this war of ideology behind us. Maybe, just maybe, it is time to pay less attention to the hype and brand-name focus and really look at the benefits and weaknesses of each of the options. Look at which solution is going to have the most to offer you and your organization. Though I do not think that there is a one-size cloud that fits all, I will still go out on a limb to predict that sometime in the future, hybrid clouds will be the standard. But in the meanwhile, can we put an end to this war of ideology and focus more on the complete solution?