IT environments are extremely heterogeneous, with multiple operating systems, hardware platforms, and business applications running in tandem, so it should come as no surprise that nearly two-thirds have multiple server virtualization hypervisors deployed. This brief looks to identify the pervasiveness of multi-hypervisor environments, and the key drivers behind these strategies.
Published: December 27, 2012
ESG recently conducted a survey of 440 IT professionals responsible for server virtualization and cloud computing strategies at large midmarket (i.e., 500 to 999 employees) and enterprise (i.e., 1,000 or more employees) organizations in Western Europe and North America. Among the topics covered was the usage of hypervisor technology. As seen in Figure 1, nearly two-thirds (65%) of respondent organizations reported using more than one hypervisor. These numbers skew higher among enterprise organizations, as well as those with more than 1,000 virtual machines (see Table 1).
What were the key drivers behind organizations’ decisions to employ a multiple hypervisor strategy? Nearly one-third (30%) of these organizations cite the need or desire to use different hypervisors in conjunction with different application workloads (see Figure 2). There is no IT solution panacea when it comes to supporting workloads since every business application is different, which applies at the hypervisor level. As organizations successfully achieve the initial server virtualization goals of consolidation and cost mitigation, many have struggled to expand their usage because of perceived application incompatibilities, performance concerns, and lack of knowledge. Major application vendors such as Oracle and Microsoft are successfully designing, engineering, and delivering their business applications—and underlying databases—to run on their own hypervisors. This is a potential area of risk for VMware since they do not operate at the application tier and are left to partner with application vendors. Other top reasons for employing a multi-hypervisor strategy included the reduction of licensing and software costs (28%) and/or the existence of legacy products (28%). In terms of reducing licensing costs, using multiple hypervisors gives IT organizations the flexibility to align specific workloads with the most optimal hypervisor—from both a technology and economic perspective—and the leverage on vendors to solicit better pricing options. Phasing out legacy products likely refers to earlier versions of hypervisors that were typically given away for free with the understanding that functionality was limited. While organizations may have relegated them to less strategic workloads, such as test and development environments, they may not have the time or resources to replace these outdated hypervisors, especially if they’re still performing their tasks.
Multi-hypervisor usage is rapidly being incorporated into IT architectural designs as part of comprehensive plans designed to deliver the benefits of server virtualization at levels of pricing, performance, and reliability that IT demands. In addition to providing pricing leverage, these strategies allow IT staffs to accommodate the specific needs of different applications by matching them with the most appropriate hypervisor, which is especially important to larger organizations with more diverse environments. This can serve to benefit companies regardless of their progression in the server virtualization adoption lifecycle. Those in the early stages that are still focused on basic achievements such as consolidation and cost containment find that they have choices when it comes to hypervisors and can avoid some of the investment premiums they might have had to previously maintain, without sacrificing functionality. Progressive server virtualization users, on the other hand, are likely to look further up the IT stack at how to manage a multi-hypervisor approach by enabling visibility, provisioning, and monitoring from a centralized management tool.
This research proves valuable for IT organizations attempting to determine how to virtualize more workloads within their existing infrastructure, as well as those looking for potential cloud migration strategies that shift the underlying compute and storage infrastructure to third-party service providers. Once a workload is encapsulated into a virtual machine, the opportunities abound— virtual machines can not only be moved between different onsite hypervisors, but also migrated to public cloud services based on various attributes including complexity, performance requirements, and economic consideration.
Although IT organizations are not making wholesale shifts to other hypervisors or to the cloud, many have embraced a multiple-hypervisor strategy, and those with plans to do so should use operating system and hardware refresh cycles as an opportunity to research the wider array of hypervisor options available to them. The IT vendor community should recognize this change and shift towards products and solutions that can adapt to these solutions, while resellers should use the opportunity to engage with customers and be part of design and implementation phases.