It was a recent generator failure during a Sydney brownout that triggered disaster recovery procedures for foreign currency exchange OzForex. The global company relies on emails to conduct online business 24 hours a day, but thanks to planning it had already moved email disaster recovery provisions off-premise and onto Symantec's MessageLabs cloud.
"Staff used laptops to access emails from the DR provider," says chief technology officer David Higgins.
"When power was restored, the data was restored automatically."
"When we receive email, it goes to our provider as well. We switch over to using these services if there is a disaster. We want instant access to the data, rather than downloading it. We want to continue to run the business while restoring power."
According to Gartner, by 2014, 30 per cent of midsize companies will have adopted disaster recovery in the cloud, up from just over one percent today. Cloud-based disaster recovery offers easier, more frequent, less expensive testing, and pay-as-you-go pricing; however, it can also come with difficult security compliances, murky security service level agreements, and few proven vendors in the market, according to an April 2012 Forrester report.
According to Forrester, cloud service DR providers include EVault, Microsoft Windows Azure, Dimension Data and IBM; and vendors CommVault, Riverbed, StorageCraft, TwinStrata, and VMware.
Brisbane floods last year also made companies rethink their disaster recovery strategy, said Adam Beavis, general manager of cloud services at Thomas Duryea Consulting.
"Building a DR strategy in the cloud reduces costs. With tools such as VMware Recovery Manager you can test DR without pulling down the production site."
"But for disaster recovery to work properly, make sure business critical apps are virtualised across the board."
Victoria Legal Aid has a 100 per cent virtual infrastructure environment, said David Emmett, ICT infrastructure manager.
It uses Thomas Duryea's cloud services for DR of its online grant application system.
"While it could afford to be down for one day … an extended outage is certainly not desirable, particularly in cases of clients facing incarceration," says Emmett.
"Virtualisation of servers makes [DR in the cloud] a lot easier to do. Restoring from back-up tapes when a disaster is declared would not address data integrity across [our internal] systems. Duplicating data to a hot DR site was needed."
But it's important that business continuity is driven by the business, not IT, says Emmett.
"If a disaster is declared, (our) staff could leave the building and still access IT services from our remote sites and portable devices such as laptops and tablets."
Disaster recovery isn't just about the fires and floods, says David Duncan, CEO Global Storage, a DR cloud services provider which uses FalconStor, EMC, Dell, and VMware.
"Ten per cent (of DR) is from big bang events, and 90 per cent from smaller events. It could be data corruption or data loss due to user error or a hardware failure, or a virus attack."
Vision Super, an industry superannuation fund, had an outage around New Years Eve 2010 due to a hardware failure.
"We brought up the system and recovered data for them in less than a day. A successful DR strategy protected their business."
But successful DR solution is only as good as the testing it goes through, warns Duncan.
If the solution isn't tested regularly, there will be configuration drift, that is, "what you had in production is not the same is what you have in the DR environment. For example, you could increase the storage size in production and forget to do that in the DR environment."