If you’ve been involved in enterprise virtualization at some point in the last decade, chances are good that you were involved with VMware vSphere. If you were in a heavily Microsoft-centric environment, or perhaps one where cost was a particularly sensitive variable like higher education, perhaps you ran Hyper-V. And, in various other more specialized use cases, it’s possible that the environment leveraged something like Oracle VM, KVM, Xen, or another lesser adopted hypervisor.
In ActualTech Media’s 2016 State of Storage in Virtualization report, nearly 70 percent of the respondents to the survey (approximately 725 individuals) reported that their organization is using vSphere in some form. Only 34 percent mentioned using Hyper-V, and less than 20 percent use any other hypervisor. Looking forward through 2016 and beyond, the choice of hypervisor will become increasingly important to the overall data center strategy.
Hyperconvergence, VVOLS, and Data Services
Thanks to virtualization, storage is becoming more integral to overall data center design than ever. With different options regarding hyperconvergence being available from different platforms, VM-level storage awareness being offered by technologies like VMware’s Virtual Volumes, and the overall data services provided by the array being a critical part of the data center design, there’s a single common denominator. Which hypervisor is chosen to run the workloads has a major impact on the options, features, and functionality available in the data center. For instance, choosing Hyper-V as the hypervisor would remove the ability to leverage VVOLS. Or choosing vSphere to run the workloads may cause one to forfeit certain technologies that a vendor has designed specifically to work with Hyper-V.
Another example of the way this hypervisor choice plays out is in kernel-based hyperconvergence. If the design calls for a hybrid or all flash array to provide Tier 1 storage and an integrated SDS platform like VMware VSAN to provide Tier 2 storage, the hypervisor matters. This design isn’t attainable with every hypervisor on the market (although a VSA-based approach could work). In the end, the hypervisor dictates the possibilities of a data center architecture more than almost any other component today.
Some IT organizations have discovered a way around this problem. They use the right hypervisor for the job each time, and are comfortable and competent managing multiple hypervisor breeds at once. In the aforementioned report, a full 35 percent of respondents report using multiple hypervisors. In fact, 8 percent of them use three or more hypervisors!
What they may look like is vSphere at the production site and Hyper-V at the DR site. Or it could mean Oracle VM for production workloads, but KVM for some specific OpenStack-based work. While this can be a bit more of a chore to manage, it opens up possibilities that those organizations running a single hypervisor forfeit.
Regardless of the size or focus of the data center, it’s more important than ever to consider the impact that the hypervisor of choice has on the overall data center design.