No, that’s not a typo in the headline. You’ve heard of virtual sprawl already, no doubt, and you may have experienced it in your own company. But virtual stall or VM stall refers to a concept coined by CA Technologies, from a landmark white paper published in August 2010.
Essentially, the idea is this: Many enterprises have launched their boats already, and are heading toward a virtualized data center or a reputable hybrid cloud solution. The problem is, they’re not halfway there yet, and the reasons are often more political than technological.
The broader topic of virtualizing large business workloads is the topic of a panel I’ll be moderating today at 1:00 pm ET / 10:00 am PT, with VMware Senior Systems Engineer Stephen Shultz and Intel Mission-Critical Data Center Strategist Mitchell Shults. (You read right, Shultz and Shults.) It’s a live chat, and I’d like to have you join in, especially with respect to this very topic: how to realign the organization around retooled, virtualized resources.
Yesterday, I talked about the problem of departmental silos that fail to align with the resource pools and shared servers that are the by-products of the virtual data center. Whereas a few years ago, certain managers were in charge of particular boxes, today those boxes don’t physically exist. But their jobs do.
“Virtual stall” is a bigger problem, which I discussed in greater detail in a story earlier in the summer. It’s more of a political/technological mashup, or what we used to call in the 20th century “sociology,” where issues that crop up in the business of migrating or transitioning systems lead to breakdowns in communication, which lead to the inevitable problems of whom to blame.
Here’s an excerpt from the CA white paper from Mason Bradbury and Andi Mann:
“Even some IT departments are skeptical of moving key systems to virtual environments for fear of the ‘guilty until proven innocent’ phenomenon – in which the virtualization infrastructure is blamed for problems until it is definitively shown otherwise. As resources are delivered to applications in pools, application specialists and others outside of the systems management department have far less visibility than in a traditional physical environment into the systems on which their applications run. When problems arise, employees are far more likely to blame the virtual infrastructure, as they cannot troubleshoot it themselves. Even if the problem has to do with the operating system or application itself, virtualization administrators worry that application owners will blame them until they can locate the problem and show that it is not a fault in the virtual infrastructure.”
Since that white paper was published, other firms including Symantec have followed up with new research. For example, take the case of the admin who happens to be the script writing expert in the IT shop. This is the go-to person for ad hoc issues that crop up, and who eventually becomes credited with finding the whirlwind solutions to every virtualization transition issue that admins fail to address. If the need for solutions were to, say, go away, then the level of appreciation they receive as a result would also subside.
As the Symantec report published last June relates, some IT shops get hung up on never-ending remedial projects that require one-time script writing — for example, “root cause analysis.” There may or may not be actual analysis involved, but the process becomes self-sustaining, taking on a life of its own. And Symantec has started measuring the side-effects in terms of dollars.
As Symantec’s Jennifer Ellard put it, “The admins simply delay virtualization which costs them increased operational complexity and inability to consolidate physical servers and end up having to pay for the energy and space utilization.”
So here’s the problem: In environments where virtualization deployment takes place in stages, the workloads that businesses would most readily classify as mission critical – the big databases, the e-mail servers, messaging, collaboration, ERP — get pushed to the bottom of the agenda. And there they sit, waiting for the political issues to resolve themselves further up the ladder.
We’ll be discussing the entire topic of mission critical workload virtualization this afternoon in the RWW live chat. I hope to see you there.