Research argues overconfidence in disaster recovery is ‘common and costly’
A new UK study from cloud disaster recovery provider iland has found that 95% of respondents have faced an outage or data loss in the past year – with 87% of that number saying it triggered a failover.
The survey, conducted by Opinion Matters and which specifically polled 250 UK decision makers responsible for their company’s IT disaster recovery plans, also found that of the 87% who had executed a failover, while 82% said they were confident it would be successful, 55% encountered problems.
The majority of ‘disasters’, as noted by respondents, were more likely to be mundane system failure (53%) or human error (52%) compared to cyber attacks (32%) and environmental issues (20%).
42% of those polled said that if their systems were down for seconds, it would be ‘highly disruptive’ or ‘catastrophic’ for businesses. Naturally, the numbers go up when it comes to minutes (69%) and hours (90%), but only 27% of companies said they were able to recover all their systems immediately following an outage.
98% agreed their systems could be available within 24 hours – clearly not a good enough standard. “What we see a lot of the time is customers are coming in with old solutions where they are getting more hours and days from a recovery time perspective,” Sam Woodcock, iland principal solutions architect, told CloudTech.
Another interesting facet of the research showed the key differences between cloud-based and on-premise disaster recovery (DR). Respondents were more likely to spend more on their on-premise kit, but expected less downtime with it when compared to a cloud solution. Perhaps not surprisingly, iland’s message from the study is that cloud DR can be both cost-effective and minimise downtime.
DR testing was also examined; 37% of respondents are testing infrequently or not at all, while 63% has a trained team that tests DR either quarterly or twice a year. The study argues that more can – and should – be done in this regard.
“Traditionally, a long time back solutions were quite intrusive,” said Woodcock. “Some solutions involved replication actually being paused – reducing your recovery point objectives. That’s obviously not beneficial to the business if a disaster were able to hit during that test – potentially there would be some additional data loss with those more traditional solutions.
“What we’re seeing more, as we’re progressing through, [are] products come to the market that give you the ability to test non-intrusively,” he added. “That’s really giving an upward trend into the frequency that people test, and how frequently people test as well.”
iland gave three recommendations for organisations to get the most of their disaster recovery plans. Companies need to ensure a balance between downtime and cost, ensure testing is easy and cost-effective, and ensure that their DR solution helps maintain IT compliance.
- » The cloud migration landscape: Multi-cloud and hybrid battle for supremacy as security remains key
- » A guide for database as a service providers: How to stand your ground against AWS – or any other cloud
- » The five key things every executive needs to know about identity and access management
- » Enterprise NoSQL adoption is now mainstream: What will happen from here
- » Financial services moving to hybrid cloud – but rearchitecting legacy systems remains a challenge