Virtualisation and Disaster Recovery

Virtualisation may not have all the answers when it comes to disaster recovery, but it can do things that basic tape or online back-ups cannot. It makes it easier to accomplish the three mandatory parts of a successful recovery: restoration of the data, the application using the data and the operating system required to make the application work. Tape or online back-ups may give you your data back, but if the servers have gone, then applications need to be installed all over again – and so do the operating systems with their passwords, permissions and configuration specific to the company concerned.

With virtualisation software such as VMWare, physical machines can be separated from operating systems. More than one operating system can run on one system; either different versions of the same operating system or different systems, such as Microsoft Windows and Linux. Servers with only one operating system typically show CPU utilisation of between 5 and 15 per cent. In other words, there’s often room for simultaneously running two or more operating systems or “virtual machines” on the same server. With multiple servers and the right modules in place, automated, “hot failover” configurations are possible for virtually instantaneous disaster recovery.

Is virtualisation for everyone?  Some applications are less well suited to virtual environments: for example, large database applications may make it necessary to dedicate a server to get the requisite level of performance. For the others, there’s a calculation to be done. On the one hand there is the cost of additional software licences and staff training and on the other the more efficient use of fewer servers. However, if one of those servers fails, the different environments running on that machine have to be made available elsewhere, meaning building in spare capacity – an ongoing cost to be compared with the impact of an interruption to the business over a longer period.