Flexible, fluid, and fast-evolving, software has regularly outstripped the capacity of the hardware it runs on, which by its nature is more rigid and slower to change. For most computer users that difference might mean occasionally sluggish performance or an inability to use certain features, depending on what hardware/software combination they're using. When the software intelligence in question is virtualization, however, the stakes are much higher.

Since virtualization's foundational benefit is to operate multiple application servers on each physical server, virtualization gradually reintroduces single points of failure back into the IT infrastructure years after server-based architectures minimized them. Distributed server architectures resist complete system failure by running critical applications on discrete physical servers. Server crashes could only take out one or two critical resources at a time, which is why distributed environments could tolerate servers' low reliability compared to the mainframes and minicomputers they replaced. The ensuing data center sprawl, however, brought server architectures full circle through virtualization. With five, six, seven, or more critical applications and databases on a single server, virtualized IT environments are even more vulnerable to catastrophic crashes than mainframe-minicomputer environments because they have the same concentration of resources but not the same hardware reliability.

None of this is any reason to shelve virtualization strategies, of course. Virtualization's benefits are too significant to ignore. Corporate IT managers have aggressively embraced virtualization to rein in data center sprawl. Almost 80% of companies surveyed by the Yankee Group are using some kind of server virtualization technology, and 85% of the money spent on virtualization software went to server consolidation projects. Implementing several virtual servers on a single physical server raises utilization rates from an average of 10% per box to as much as 80%, according to the Yankee Group's 2007 report on server virtualization. Virtualization's entrance didn't come a moment too soon, for image-conscious companies, either. Quoting a McKinsey & Co. study, the New York Times predicted that at their current pace, data centers will be the biggest producers of greenhouse gasses by 2020.

To realize its full potential, virtualization must go hand-in-hand with reliability strategies that consider its few weaknesses and address them with the right hardware and management tactics to reduce them.

No Built-in Virtual Reliability
Reliability and business continuity aren't front-of-brain in a virtualization discussion. Part of that omission can be attributed to virtualization's nature. The ease of setting up multiple application servers in a virtual environment can make virtualization look like its own reliability strategy. Implementing backup copies of applications on virtual servers means there's a copy in place and ready to go in case of a server crash. With backups ready to go, IT need never be caught flat-footed in the event of an application crash.