Try as it might, the enterprise is all but destined to encounter a multi-hypervisor environment, either at home or in the cloud. And that adds a new wrinkle to the management challenge as each virtual platform has its own set of unique properties that make interoperability difficult at best. All is not lost, however, as new generations of management systems foster federated virtual environments that at least provide a single pane of glass to manage disparate environments as a single entity. As Embotics CEO Jay Litkey puts it in a conversation with IT Business Edge’s Arthur Cole, the idea is to normalize the differences between platforms so the enterprise enjoys a more workable data ecosystem.
Cole: VMware has had a good run in the data center so far, but it seems more organizations are warming up to Hyper-V of late. How do multi-hypervisor deployments complicate the virtual environment from a management perspective?
Litkey: Multi-hypervisor deployments complicate the virtual environment from a management perspective because of the inherent silo created by each platform. Each hypervisor has its own paradigm, user interfaces, terminology and out-of-the-box management capabilities that need to be learned by system administrators.
When more than one hypervisor is deployed, administrators are then forced to context-switch and mentally toggle between the differences of each platform as part of managing each. One simple example is that VMware has its vCenter console and Microsoft has its System Center Virtual Machine Manager (SCVMM). Two very different tools that one administrator needs to learn in order to manage and scale up either platform.
Cole: Is it necessary to foster interoperability between hypervisors in order to share workloads, or should they be viewed as distinct virtual architectures? And if it's the latter, don't we run the danger of replacing one silo architecture for another?
Litkey: If one is to simplify the management of multi-hypervisor environments, then federated management using an abstracted single-pane-of-glass, like Embotics vCommander, becomes important. That way, a set of standardized administration tasks can be performed across one or more hypervisor platforms while effectively normalizing differences in paradigms, user interfaces and terminology.
Interoperability and portability of workloads between hypervisor platforms also knocks down silos from building, and a user retains his right to exercise freedom of choice as time goes on, based on constraints that matter most to them, such as performance, cost or reliability.
Cole: Ultimately, though, won't most enterprises need to deal with multi-hypervisor environments at some point as more and more workloads are finding their way to the cloud?
Litkey: Yes, most enterprises will need to deal with multi-hypervisor environments because history will repeat itself. One only has to look at data center operating systems like Windows, Linux and Unix, or databases like Oracle and SQL, and how many IT organizations currently have mixed environments to support the diverse needs of their businesses.
Hypervisors and clouds are no different in that they will differentiate in the market and be optimized to best-fit varying usage scenarios. Some will be cheapest, others will offer premium performance, while others still may be the most secure. Freedom of choice and the ability to optimize as time goes on will only grow in importance.