The advantages of managing DR from the hypervisor rather than the storage layer

Zerto's Chris Snell explains why DR/BC and replication need a rethink in the age of hybrid cloud

Disaster recovery has always one of those things that everyone needs but which few mput enough time into thinking about, favouring the default option, which may not always be the best choice. In the current world, which is defined by hybrid and private clouds - and above all virtualisation - the entire model of DR/BC has changed, but people may still be reluctant to think about it, preferring to stick with the old ways of doing things.

"Our customers tend to be those with private and hybrid clouds," said Chris Snell, solutions engineer at Zerto. "They want replication of private clouds, or DR as a service in hybrid clouds, or in the public cloud such as AWS."

Historically, he said, customers were locked into the big storage vendors and were effectively forced to use those vendors' DR solutions, but in the new world this replication is often on the wrong layer. If the IT team wants to restore individual files of VMs this can be an arduous process.

Zerto has moved away from storage-based replication and into hypervisor-based replication. "We play inside the hypervisor itself so there's no lock-in," said Snell.

He went on to say the company's technology does not use snapshots, and as it works in the background there is no disruption to the working environment.

"We allow you to replicate and recover into a DR target in an automated fashion."

Enterprises demand very low losses of data and very quick recovery times, he went on, saying that on average with hypervisor-level recovery they can expect to lose just 10 seconds worth of data and that a complete recovery can take two minutes.

Synchronous data replication is unnecessary for the majority of data, so the money spent on expensive licensing might be better used elsewhere. Also synchronous replication carries the risk that if the data is corrupt then the replica may be corrupt too, with no way of rolling it back.

"By automating disaster recovery and making it continuous and simple, taking it away from the storage layer, we've had cases where even non-IT people have been able to do a recovery with a few mouse clicks," Snell explained.

Because the DR/BC is done at the hypervisor level it is storage agnostic. There can be knock-on benefits of this approach, Snell said, including enabling simple offsite backup to the cloud, or to another data centre.

It can also help firms to sweat their assets. When they upgrade their storage, older equipment can be moved into the DR site with hypervisor-based software able to bridge across the new and the old. The same technology can also be used to make data migrations much simpler.

Zerto was a sponsor of the Computing Data Centre Summit