As companies advance and progress with virtualization deployments across the newly formed virtual or dynamic data center, they typically do so across what can be defined as a virtualization maturity model with a goal of delivering applications and resources to users in a reliable, secure, and cost-effective manner.
Early on in the maturity model of virtualization, phase one or cost savings, many companies struggle with a concept that has become known as "VM sprawl." The term "VM sprawl" was coined because once an administrator or user became familiar with creating virtual machines, they quickly started rolling them out one after another due to their ease of deployment. Where a physical machine could take hours or even days to deploy depending on the processes and procedures wrapped around it, a virtual machine typically takes only a matter of minutes to copy before it can be rolled out and spun up for use. With it, virtual machine sprawl ensued.
So while virtual machines were supposed to help with consolidation, ease of management, and cost reduction, VM sprawl created severe problems of its own, negating many of those goals and causing such problems as performance issues, wasted resources, runaway costs, security issues, and more.
The companies that met this challenge head on and kept VM sprawl in check were able to proceed past the first phase of the virtualization maturity model. For some, this was enough. For others, they would need to move beyond the low-hanging fruit in order to achieve more than just the vendor-promised cost savings that came with simple server consolidation. The next phase in the virtualization maturity model would delve into production readiness with mission-critical applications.
However, as companies moved from one phase to the next, they also went from dealing with VM sprawl to a new challenge: VM stall.
Andi Mann, VP of product marketing for virtualization at CA Technologies, has been talking about VM stall for a while now and defines it as essentially where virtualization rollouts seem to slow down or even stop as new challenges force IT to reconsider their virtualization strategy -- typically after 20 to 30 percent of servers have been virtualized. Without a majority of servers virtualized, IT is likely to miss out on key virtualization goals.
"There is no one cause, I believe, but there is a common pattern," said Mann. "Typically, organizations go 100 miles an hour at the low-hanging fruit of virtualization, primarily for server consolidation. In this phase, they consolidate simple, homogeneous, low-priority, single-server, Tier 2/3 applications and services -- like development and test systems, file servers, print servers, Web servers, departmental applications, internal service apps, etc."
Mann continued, "However, once these are virtualized, many organizations strike a 'perfect storm' of challenges that slows their virtualization rollout, or stops it entirely. Some causes at this stage include greater complexity of services and applications, higher demand on scarce virtualization skills, limited visibility into a growing deployment, increasingly heterogeneous systems, and greater resistance from risk-averse application owners and recalcitrant application vendors."
Like Mann, Dave Bartoletti, a senior analyst with the Taneja Group, agrees that VM stall is alive and well. With his clients, Bartoletti consistently hears that VM stall exists somewhere between the 22 and 25 percent virtualization rates, but he believes this is a natural and expected situation.
"We're exiting the first wave of consolidation and the low-hanging fruit has been virtualized," explained Bartoletti. "If virtualization only delivered consolidation, we'd be done. But it's moving beyond that use case quickly, as customers leverage virtualization for so much more: application mobility, better data protection and DR, branch office consolidation, desktop efficiency, etc."
Bartoletti continued, "The second wave of issues is always harder when a core technology matures. Server virtualization essentially paid for itself in CAPEX savings, but when we virtualize Tier 1 business-critical applications, or user desktops, CAPEX savings take a backseat to application performance and IT efficiency, and this is why we're stalling."
We've identified that VM Stall typically takes place somewhere between 20 to 30 percent virtualized -- so how does one get "over the hump"?
When asked, Mann said that IT needs to revisit its approach to virtualization. It must move past the project-based approach and look at virtualization more strategically. It must provide visibility by discovering, mapping, and rationalizing existing virtualization deployments.
To do that, Mann listed the following thoughts:
- Existing virtualization architectures should be re-evaluated, especially management technologies and approaches.
- IT must be prepared to not just prove, but guarantee service quality, in order to satisfy mission-critical, Tier 1/2 application owners.
- Manual processes must be automated to free up skilled resources and eliminate errors.
- People, process, and technology management must be more standardized to allow virtualization deployments to scale.
"I think we're in the same place as we were with Internet/intranet technologies 10 years ago," said Bartoletti. "The easy, public stuff went up on the Internet or on a corporate intranet without much hassle. And it took a few more years, and some trial and error, before companies felt confident enough to do business on the Web, or allow Web access to sensitive data, or replace a legacy desktop application with a browser-based version."
He went on to say, "The current hot trends in virtualization all have a direct impact on both data security and user experience, two things no IT department messes with lightly. And I expect we'll stall for a few more years until the vendor ecosystem matures, at which point we'll wonder why we ever hesitated to virtualize everything."
Where are you on your virtualization maturity path? Have you broken through the 20 to 30 percent glass ceiling?
This article, "VM stall: Breaking through the second phase of virtualization" was originally published at InfoWorld.com. Read more of David Marshall's Virtualization Report blog and follow the latest developments in virtualization at InfoWorld.com.