Improved management was one of the chief selling points for virtualization several years ago, but is it possible that perhaps it was just a bit oversold?
The more virtualization becomes part of the workaday IT experience, the more it becomes clear that the management of virtual systems is every bit as complicated, if not more so, than management of physical infrastructure. At best, you can argue that virtual environments lend themselves to greater flexibility and the introduction of new capabilities, but not a simplified management regime.
Part of this is due to the learning curve that every IT shop experiences once the initial virtual layer is established, according to author and consultant Charlie Schluting. Everything from deployment to monitoring and tuning must be relearned in the virtual world, which wouldn’t be too much of a burden if not for the way virtualization starts to take its toll on physical resources as more virtual machines come online. At this point, the main burden shifts from constant provisioning and decommissioning to load management and auto-migration policy maintenance.
If you play it right, though, you can counter those tasks by consolidating once-disparate systems so they can be managed as one environment. A good example is backup and recovery, says Elias Khnaser of systems integrator Artemis Technology. By treating DR facilities as simple extensions of the data center, you can better leverage existing virtual management systems, have a set of resources ready for off-loading data during peak periods and routine maintenance, and have a fully functioning backup facility in case disaster strikes.
Another problem in the virtual management arena is that many CIOs are trying to apply the same policies and procedures of the physical world to the virtual one. One key element that is missing in many virtual deployments is advanced analytics, according to Dave McCrory, founder and CTO of Hyper9 Inc. Without the right data on such items as capacity, configuration, performance or usage, it’s almost impossible to gauge what the problem is and how to fix it.
And in storage circles, at least, there seems to be a tendency to keep the focus on direct resource management when the new virtual environment demands a shift toward application-based management, according to analyst John Webster of Evaluator Group. Now that many data management services have been transferred from the storage controller to the OS, you have the ability to view both the application layer and the storage layer from a single vantage point. Couple that with new traffic processing chips and fiber optic sensors and you have an environment where application needs can produce automated responses from storage — no need to constantly monitor and reconfigure your storage environment because it will find all the information it needs right on the application-seeking resources.
That kind of automation certainly would cut down on a lot of busy work, but let’s be clear: Automation requires a fair amount of human oversight as well, just to be sure that everything is working the way it’s supposed to.
There will always be plenty to do around the data center. When evaluating virtual management systems, the thing to keep in mind is ensuring that your human resources are devoted to the most productive tasks.