Business back-up procedures

Recent events have put the issue of business continuity and disaster recovery at the forefront of many CEOs' minds. All businesses should prepare for the worst, be it the total loss of their premises, computer failure or simple human error; losing the information stored on systems can be catastrophic.

Mike Osborne, operations director of ICM Computer Group, believes that a shift is occurring away from magnetic media such as tapes towards synchronous and asynchronous models of data back-up as the latter become more affordable. The vulnerability of tapes was highlighted last month as the retail finance division of Citigroup admitted to misplacing tapes containing information on four million customers.

Related: Why do so many CEOs lack a business continuity plan?

Synchronous technology is the best option, but also by far the most expensive as it needs particularly sophisticated technology. Basically, it backs up all work in real time, meaning that every time a change is made it is simultaneously duplicated on a system off-site.

This means that, in the event of a disaster, no data should be lost and there should be minimal delay in recommencing business. ‘The time it takes for your staff to get to the off-site premises is the time it would take to get back up and running,’ says Osborne.

However, at around a tenth of the cost, the asynchronous model is what most businesses will choose unless it is absolutely vital to not lose any data at all. Under this system, changes are backed up periodically, generally at the end of the day, and then sent, via the unused broadband line after work has finished for the day, to an off-site location.

‘Of course, with this, some information is likely to be lost if a recovery had to be made,’ warns Osborne. ‘How much depends on how frequently the back-up occurs and the time of the disaster.’ For example, if the back-up takes place once a day at 9pm, but the office systems are lost at 7pm then no changes from that day will be saved.

With tapes, even assuming they have been stored off-site (Osborne believes a significant number of businesses store tapes in the same premises as their main servers) and are therefore safe from a building disaster like a fire, it takes 24 to 48 hours to retrieve all the data.

When thinking about back-up, Osborne suggests looking at the ‘critical data. Do you need everything backed up or perhaps just one application, email, for example? This has a major bearing on costs.’

Speaking of costs, he estimates synchronous back-up to cost between £500,000 and £1 million. Currently, it is only really worthwhile for organisations such as banks, although as costs fall, more firms will take it up, he predicts. The expense reduces considerably if just one application is backed up, perhaps to £30,000 a year.

At between £10,000 and £50,000 for a full back-up, asynchronous systems fall more into most firms’ price brackets.

Whatever system is being used, Osborne strongly advises regular ‘test’ retrievals to ensure everything is working correctly. ‘It would be a real disaster if the worst happened and you discovered you hadn’t actually backed anything up after all.’

Case study – Recovery put to the test

In March 2004, a fire in a BT communications tunnel affected 130,000 business and residential phone lines in Manchester, including those at the offices of Yell Group’s telesales department, who were unable to dial out and therefore ply their trade. Brian Davies, the service continuity manager, immediately contacted ICM to warn them that Yell would be invoking the recovery plan.

‘As luck would have it, we had been testing the recovery plan just the week before,’ recalls Davies, ‘so the actual recovery was pretty much a blueprint of the test. We run a test for each of our locations every year and try and get as many staff as possible to go through it for realism’s sake.’

This attention to detail paid off as within hours the hundred or so telesales staff had been transferred to the recovery centre, where the computer network was set up and the relevant data, backed up from the regular office, was accessible. Some hard work through the night saw the sales team operational again the following morning, meaning only one working day had been lost. And, as it turned out, the team met their weekly targets anyway.

‘It’s important to remember that the recovery centre is functional – it’s not going to be plush offices,’ he adds. ‘We’re talking about business survival, not business as usual.’

Marc Barber

Marc Barber

Marc was editor of GrowthBusiness from 2006 to 2010. He specialised in writing about entrepreneurs, private equity and venture capital, mid-market M&A, small caps and high-growth businesses.

Related Topics