In addition to protecting your data by keeping copies elsewhere, you can protect your productivity by keeping copies of your critical servers. This is known as a “Hot Backup” or “Hot Backup Site”. Not many companies do this today as it is simply too expensive in equipment and man-hours to maintain.

A compromise is something called clustering, where you still need duplicate server hardware, but you can share the storage (usually via SAN or NAS) and other infrastructure. It is still fairly complex to maintain, but it does provide 24x7x365 availability or very close. Clustering allows the IT staff to do maintenance and upgrades on part of the cluster while the business runs on the other cluster servers. Large-scale computer system manufacturers like IBM (Power) and HP (HP-UX) built resource sharing/partitioning for large systems to help facilitate cluster and cluster-like operations.

Virtualization can trace its lineage back to these partitioning roots (and even back to time-share days) with perhaps the first mainstream example being Macintosh users who wanted to be able to run Microsoft applications. By 2005, there were several virtualization options on the market that didn’t just split CPU and memory resources into fixed partitions, but could let virtual machines shared resources dynamically as organizations realized that they had purchased a lot of very expensive equipment which was standing-by ready to perform a critical task, but was more-or-less idle 95% of the time -- this also consumed a lot more power and more data-center floor space (the most expensive per square foot). Buying new hardware to run these services virtually was expensive, but the expense was offset multiple times by the savings.

It took a few tries, but eventually they came up with a suite of virtualization options for use from international corporations down to a lone programmer working in his basement. Today, a small business might use Red Hat’s KVM, Oracle’s Virtual Box, VMware’s ESXi, or Microsoft’s Hyper-V to consolidate systems that can all be hosted on a single set of hardware.

The redundancy advantage to this is that if your host server(s) are protected by redundancy, all the virtual machines hosted are also protected.