Virtualisation has helped to make disaster recovery and business continuity planning easy and affordable for even the smallest companies, says Warren Olivier, regional corporate account manager at Veeam Software – but many still aren’t taking advantage of the new opportunities the technology has brought.
“There are still many companies out there whose disaster recovery planning is limited to someone taking a backup drive home every night,” says Olivier. “But restoring their business from a backup like that could take hours or days. If those are hours or days they can’t afford to lose, they should seriously investigate a disaster recovery service.”
Offering disaster recovery as a service has only become possible in the few years since virtualisation has gone mainstream, Olivier adds.
“If your IT environment is purely physical, you have to restore it to identical hardware. For a service provider this would have meant maintaining separate equipment for every client, which could never be economically viable. That problem disappears with virtualisation – you can restore a virtual machine to any suitable physical equipment. So suddenly, it makes business sense.”
To bring the costs down even further, notes Olivier, “a provider might even use older or lower-specced hardware – it’s a useful way of recycling old machines that are no longer suitable for your production environment”.
The idea of using older or slower machines raises what Olivier says is the critical question for anyone considering using a disaster recovery service: What level of service do they need?
“In an ideal world of infinite budgets,” he says, “of course we’d all like to have a second building standing by, complete with a replicated data centre and all the furniture and PCs we’d need if the building burned down, or flooded, or there was an extended power outage. But that’s an enormous expense that only a very few can afford – and equally few really need.”
Getting value for money requires an honest assessment of a company’s needs, says Olivier.
“You need to set realistic recovery point objectives (RPO) and time-to-recovery objectives (RTO) for every part of your operation. Your HR department or development team might be able to cope with a couple of days of downtime, your production environment with no more than a few minutes. The higher the level of service you need the more you will pay for it, so it’s worth spending time on getting these calculations right.”
Clients should also ensure their service level agreements with a provider cover critical questions like how much of a performance drop is acceptable, whether they will accept a multi-tenanted solution where they share capacity with others, how often disaster recovery will be tested and what reports they need.
“It can get very complex,” says Olivier.
“For example, you need to decide to what level you want to test – do you just need to be sure that a server boots up, or that individual applications will still work properly? You need to think about what happens to your e-mail, to IP addresses, to your CRM system – disaster planning can be a real headache. For anybody who can’t afford dedicated in-house resource, it’s best left to the experts.”