For every IT issue that virtualization solves, it creates another problem. And after a wave of virtualization adoption, the new problems presented by virtualization are starting to take their toll on adoption.
Andi Mann, vice president for virtualization management marketing for CA Technologies, says that as “virtualization stall” continues to occur he expects to see a period of digestion as IT organizations step back to assess how they will rise to meet the management challenges associated with large-scale virtualization.
Those challenges include how to manage virtual machines as they dynamically move across an enterprise network; coping with images that tend to dramatically increase in size over time; concerns about performance limitations on servers with limited amounts of memory; and ongoing resistance to virtualization among owners of applications that prefer to see a one-to-one relationship between their applications and the machines they run on.
Mann adds that many IT organizations are also uncomfortable with not only the lack of visibility they have into the way virtual machine software operates within their environment, but also the fact that virtual machine software introduces a layer of software that cuts off visibility between the application layer and the underlying infrastructure.
Nevertheless, server savings continue to drive interest in virtual machine software. But as Mann notes, without an effective way to manage those virtual machines, the cure can easily wind up being worse than the disease as virtual machine sprawl starts to choke the enterprise.
Mann says that the only effective approach to the problem is more structured approaches to IT management based on guidelines such as the IT infrastructure library (ITIL) specification and more reliance on IT automation tools. Where you once might have had one IT administrator for 10 servers, you’re now looking at a scenario in which one IT administrator may have to manage 100 virtual servers. No IT organization can keep pace with that kind of growth without some increased reliance on IT automation.
In the meantime, Mann says he expects to see a period of virtualization stall through most of the this year as IT organizations try to figure how to take a more strategic approach to managing virtualization, as opposed to a tactical approach that solely focuses on reducing server costs without appreciating the total cost of virtualization ownership in the enterprise.
Click through to see key issues that are getting in the way of a good idea.
Sometimes it’s just simpler to buy another $2,000 server than master virtual machine software.
Most IT organizations are not equipped to manage hundreds of virtual machines.
While many new servers can run 20 or more virtual machines, most existing servers can handle about four to six.
There’s not enough skilled talent to keep pace with virtualization demand.
Despite proof to the contrary, many applications owners don’t trust virtualization performance.
Many application vendors still license software on a per-processor basis.
Network managers are wary of the complexity of virtual machine software that can move at a moment’s notice.
The longer a virtual machine is deployed, the larger the images start to become.
Most storage resources are allocated to specific applications.
Any time something goes wrong, the new software is blamed regardless of what happened.