James Thew - Fotolia

Support for multiple hypervisors could solidify VMware's future

While VMware has experimented with multi-hypervisor management before, it should get serious and work it into vCenter Server to stave off rivals.

In the beginning of x86 virtualization, it was VMware or nothing when it came to hypervisors. Now with multiple hypervisors gaining in capability with each successive release, it's time VMware baked support for these competitors into vCenter Server.

I think this is where the real opportunity lies for VMware. In the past, VMware has played with multi-hypervisor management as a pet project. It has released free plug-ins that allowed some basic management of Hyper-V hosts. This was nothing more than a science project. It's time for VMware to take multiple hypervisor management seriously, build something real and use it as a way to stay relevant in competitive spots where vSphere is likely to lose.

The hypervisor market is seeing increased competition. There has always been a little pressure from Microsoft while others have been fringe use cases for most. But Microsoft is not going to continue to swing and miss with Hyper-V; eventually it will build solid momentum and start to hurt VMware where it counts -- in the wallet.

Linux-based hypervisors, such as KVM, are also getting some notice. In the past, KVM was strictly for Web shops, service providers and Linux fan boys. But after gaining some traction, now the small and medium-sized businesses and enterprise customers are taking notice. With the growing popularity of Web-based apps and OpenStack hype, more organizations are considering open source as one way to reduce licensing costs.

Lastly there is the XenServer crowd, which is pretty much isolated to select Citrix installations. Recent performance, storage and vGPU updates to XenServer have shown there may be some life left in those binaries.

The opportunity is there

VMware has built something special with vCenter Server as the management layer for the vSphere hypervisor. I would say that it's the premier management choice for a hypervisor -- as long as that hypervisor is vSphere.

If you are a Hyper-V customer, you use System Center Virtual Machine Manager (SCVMM) no matter what the scale. If that is the case, then you are either laughing and/or cursing because SCVMM is widely criticized as being a terrible management product for Hyper-V. It suffers from poor performance and is confusing to learn, to name a few of its issues. SCVMM alone might be a solid reason why more organizations have not yet deployed Hyper-V as a replacement to VMware. While Hyper-V as a core hypervisor has gotten a lot better, the management is where it's lacking.

Citrix has built XenCenter as the management layer for XenServer. I've never heard anyone speak highly of XenCenter. Adding management support for XenServer to vCenter Server may not be at the top of VMware's to-do list, but it could be a way to draw customers away from Citrix if a superior administration tool was available.

Lastly if you are deploying KVM, there are not a lot of management options available for you. The core hypervisor product has no management product other than the command line. This is not attractive to the majority of customers who are not already a large Linux shop. The one solid option is to purchase and deploy Red Hat Enterprise Virtualization Manager as the hypervisor management offering. This might makes sense if you are a shop that pays for Red Hat license, but if you use CentOS or other distribution, you probably don't want to pay for Red Hat management.

Start with KVM

VMware can build real multi-hypervisor management into vCenter Server and make it the de facto hypervisor management choice for everyone. The first on the list should be the KVM hypervisor. It has the biggest opportunity and won't have the direct competition it would with Hyper-V. Also, it would fit nicely with its open source projects around NSX, OpenStack and others.

Admins would be able to manage vSphere and KVM hypervisor clusters from a single vCenter instance. The goal would be to build as many of the vSphere-type features as they could within vCenter Server for KVM. This would make certain aspects easier, such as creating and managing KVM clusters.

The partnership would make it easy for admins to create, clone, delete and manage virtual machines without touching the command line. If they could build in the ability to convert VMs between the two hypervisor platforms, it would be another big benefit. This would allow organizations to run vSphere in production and KVM in other environments, and migrate workloads to production when ready without a complicated process.

I realize this is probably a long shot, but VMware should take multiple hypervisor management seriously. This would put the company in good standing with more organizations and allow them to pull away customers who are using KVM and other hypervisors.

Dig Deeper on Creating and upgrading VMware servers and VMs