The cutting edge of virtualization technology may have set its sights on virtual PCs, unified network fabrics and other esoteric applications, but server consolidation remains the primary driver for most data centers.
In fact, only a handful of enterprises have begun the process of virtualizing their server farms, according to most recent surveys, although the pace is likely to pick up as energy costs and competitive pressures drive organizations to increase performance even while paring down their hardware infrastructures.
But as those who have already taken the virtual plunge have no doubt realized, consolidating servers is not just a simple matter of powering up the virtualization layer and then pulling equipment out of racks. There is a long list of factors to consider with any centralization project and a wide range of land mines that need to be avoided to prevent service failures.
One of the main concerns is the resiliency of remaining hardware, according to Vision Solution's Bill Hammond. Tape backups, vault systems and even hot site backups can restore lost service with minimal downtime, but organizations requiring 24/7 service should consider building a high-availability cluster that can be switched over to in an instant. The beauty of such an approach is that organizations will already have the hardware at hand to create such a system courtesy of the servers that were just decommissioned during the consolidation.
Still another concern is establishing the proper management processes for the new virtual environment. According to GridToday, a poorly managed virtual infrastructure will do little or nothing to simplify operations or reduce costs, and could cause an explosion of new components and devices that end up reinforcing the very silos that should be integrated. Unfortunately, most available virtual management tools lack the ability to construct a complete service-based model of the enterprise.
For a truly successful consolidation, you need to forget about the hardware for a moment, according to IT consultant Brad Harris. Before the first box is unplugged, enterprises should take stock of their organizational and administrative needs. Key considerations are changes to control, configuration and admin management, workload inventory practices and a host of other functions. Once the governing framework is in place, the actual consolidation can proceed in a more organized fashion.
It's also important to remember that consolidation is not the end result of virtualization -- it's only the first step in a soup-to-nuts infrastructural change that will affect everything from storage to networking to applications, according to Jim Damoulakis of GlassHouse Technologies. You're going to need buy-in from all facets of your organization, with ongoing contributions to the integration of operational processes into existing workflows.
Nothing worthwhile is ever easy, and it would be a mistake to ready anybody's "Top 10 List of Virtualization Pitfalls" and think you have all the bases covered. Unique environments will demand unique approaches. But at least by knowing what others have encountered in the past, you can avoid those mistakes in the future.