Payroll systems are just one of the vital applications used in finance and HR departments, but their importance means that a sudden complete failure of the system - stemming, for example, from a major power cut, a fire, or even a physical attack on the premises where the office is located -would be extremely serious for any organisation.

Disaster recovery strategies exist precisely to provide a contingency against an event that is, indeed, potentially disastrous. Disaster recovery became popular as a strategic policy in the 1980s due to the somewhat belated realisation that financial institutions could be extremely vulnerable to major incidents such as fires, serious power failures and even terrorist attacks.

Disaster Recovery

The usual way to plan disaster recovery has, understandably, been to set up an alternative physical location that’s fitted out with desks, chairs, computers and other essentials. The idea is that employees can then be transferred physically to the back-up location and simply continue where they left off. Of course, even at the time, it was understood that things would never be quite as easy as that. Installing computers was all very well, but what about data held electronically at the stricken office? What about, say, payroll transactions (such as BACS transfers) going on at the moment, when the system went down? What about transactions that were about to take place?

Fortunately, few institutions have found themselves actually having to make use of their disaster recovery resources (or contingency planning, as it was often more euphemistically called), or the holes in the apparently commonsense scheme might have been more apparent.

Today, disaster recovery is more important than it has ever been. The tragic fact that there have been incidents or attacks, which have had a disastrous effect on some institutions, has placed a new spotlight on disaster recovery. This has led to questions as to just how effective traditional approaches are likely to be if the facilities really need to be used. Fortunately for disaster recovery planners, a new phenomenon in the way computers are used is offering a different approach to to tackling this problem. This may easily end up transforming the provisions that institutions make for disasters and for all kinds of threatening contingencies.

Virtualisation

The name of this new game is virtualisation. As you might expect, the word essentially refers to making physical computer resources virtual ones. What this actually means is that virtualisation looks for ways to have previously physical computer resources stored in software.

A virtual server or desktop will ‘feel’ no different to a user than a traditional physical one. An important benefit of virtualisation is that your office will have fewer physical servers, meaning there is less need for electrical power operating the hardware and for air conditioning. You would also have more space. Savings can usually also be made in contracts for hardware maintenance.

Virtualisation really comes into its own, however, in the area of disaster recovery. Why? Because virtualisation enormously simplifies disaster recovery and business continuity processes. Virtual machines are simply files residing on a host server. These files can be easily copied to another location on an ongoing basis to provide recovery from a disaster in seconds or minutes. In other words, the servers in your remote location can be running exactly the same data, functionality and transaction-making capacity as in your main office.

As things stand today, many organisations base their disaster recovery resources around expensive duplicated data centres that contain identical hardware. Typically one of the data centres will not be used at all for day-to-day activity, being employed purely as a back-up data centre in the event of a disaster. This ‘ready for action’ data centre will contain many banks of PCs that staff can use if their primary office is rendered unusable for whatever reason.

By employing virtualisation, though, organisations can enormously improve their ability to restore normal service followinga disaster. Of course, the computers, seats and tables for the staff will still be needed, nor — and this point must be emphasised — does virtualisation used in disaster recovery actually avoid the need for the secondary, back-up physical location. But the amount of computer hardware in the main office and back-up office can be reduced because so many of the servers are now virtual servers.

Snapshots of data

Above all, though, virtualisation allows ‘snapshots’ to be taken of servers at various points throughout a day, or even on a pretty much continuous basis. So, for example, as your payroll is prepared for the month in question, all the data that is part of the preparations can be continually ‘snapshotted’ and relayed by virtualisation to the remote location.

These snapshots can then be used as reference points to ‘return’ to, from which the operations can then continue as before. So, if there is a sudden interruption of service at your main office, the disaster recovery process can take over very smoothly. Little or no data will have been lost, and transactions that were in the pipeline when the interruption to service occurred will still be in the pipeline because they will have been part of the snapshot.

The very nature of disaster recovery centres is that they will only be used for occasional testing or, at the very worst, for providing alternative services to staff affected by a disaster at a main office. For this reason alone the emergence of virtualisation as a realistic alternative to physical computers should make financial sense to the majority of organisations who require disaster recovery facilities. Virtualisation can significantly reduce the cost of maintaining a seldom-used disaster recovery centre, while at the same time making the centre itself far more efficient and agile.

How popular is virtualisation right now? Estimates of the proliferation of virtualisation in major markets around the world vary, but it appears that virtualisation is growing in all areas, not just server virtualisation, and that it will continue to experience growth until it becomes a mainstream technology in, or around, 2011/2012, depending on which analyst group you choose to believe. So opting for virtualisation now will put you well ahead in the race for reduced costs and maximised efficiency.

This article appeared in Pay magazine in December 2008.

Please read more about our work in Microsoft Technologies