Technology giants HP, IBM, Intel and Red Hat joined forces last month to create the Open Virtualisation Alliance (OVA) with the aim of driving adoption of an open-source virtualisation product called Kernel-based Virtual Machine (KVM).
KVM is seen as a potential competitor to VMware’s hypervisor virtualisation product within the enterprise market. However, Red Hat, the original provider of KVM, has struggled to generate any significant market traction up until now.
“KVM is obviously late to the market when you look at the other hypervisors that have become mainstream,” says Chris Wolf, research vice president at Gartner (pictured).
“VMware’s vSphere, Citrix’s XenServer, Oracle’s VM and Microsoft’s Hyper-V are all far more popular with the enterprise than KVM is.”
The core problem with KVM, argues Wolf, is that Red Hat has failed to create a “vendor ecosystem” around the hypervisor, and consequently its competitors are attracting customers with their more rounded offerings.
“At this point enterprise customers complain that they can’t use KVM in production workloads if, for example, they can’t secure the hypervisor with a third-party security product, or if their backup product doesn’t support the KVM hypervisor,” says Wolf.
“This is now a pretty serious push to build a vendor ecosystem around the hypervisor, and start bringing products to the market that truly integrate with the KVM solution.”
Wolf predicts that such products will start to appear in the next few months, and that there will be significant activity in this area by 2012.
But how does the KVM technology, as it stands now, compare to VMware’s market-leading vSphere?
“It is getting there. It can virtualise both servers and server operating systems,” explains Wolf. “But when you get into richer mobility, dynamic workload balancing and real deep storage integration as part of the provisioning process, it is still lacking. VMware has had a tremendous head start over KVM, and KVM will never catch up in terms of features, as VMware will just continue to innovate.”
The sudden support for KVM, which is Linux-based, suggests that these major vendors are concerned about VMware’s dominance within the market.
Ovum analyst Laurent Lachal argues that the likes of IBM and HP now recognise that if Linux is to remain a competitive force against Microsoft’s dominant operating system, KVM must succeed.
“The success of Linux and the success of KVM are intertwined. Where firms have an interest in Linux succeeding, they have an interest in KVM succeeding,” he says.
Lachal’s argument stems from the rising concern in the industry that VMware is looking to replace the major operating systems like Windows and Linux with its hypervisor offering. This is because a hypervisor works as a manager of numerous operating systems, operating in the layer below the OS. At the moment, operating systems and hypervisors jointly control the hardware drivers.
“For Linux to remain relevant, it needs to remain relevant as a virtualisation platform, and Microsoft has recognised the same is true for Windows,” he adds.
However, there are concerns that VMware will develop its hypervisor so it operates the hardware drivers on its own - it currently does this with help from an operating system. As a result, the operating system would become a selection of packaged applications doing little else but running images. Traditional OS providers would see their influence significantly diminished.
“VMware is trying to push the OS aside in order to have its technology at the centre stage, and the Microsoft and Linux communities are trying to elbow themselves back into the frame by saying the operating system is as good a virtualisation platform as anything else,” says Lachal.
“Microsoft is attempting to do this with its Hyper-V solution, and now the Linux community is doing the same with KVM. The reason this alliance has been formed is that the vendors want to maintain a diverse landscape and not fall prey to a VMware monopoly.”
However, this may prove to be more difficult to achieve than Red Hat anticipates, as changing to KVM would not be a simple cost-free switch for the large amount of enterprises currently using VMware.
“VMware doesn’t have a lot to worry about. Red Hat can make a technical argument as to why customers should switch hypervisors, but the fact is that it is very cost-prohibitive to do so,” said Wolf.
“The real problem isn’t the virtual machine itself - it’s the dozens of management applications and operational software that these enterprises have tied into a particular hypervisor.
“It is disruptive and costly to move an application to a new hypervisor because there is a lot of operational management that needs to change to get it to work correctly.”
By eliminating high entry costs for big data analysis, you can convert more raw data into valuable business insight.
A discussion of the "risk perception gap", its implications and how it can be closed