Modern Infrastructure Editor-in-Chief
Published: 19 Mar 2014
Disaster recovery used to be the province of the rich and famous: Only the most deep-pocketed organizations could afford the equipment, effort and expertise needed to put together a comprehensive DR plan.
In the past decade, though, virtualization has lowered the bar to disaster recovery. By running a hypervisor at a remote site, organizations can eliminate the need for exact hardware replicas at the DR site, greatly reducing costs and overhead.
Now, the public cloud has pushed DR one step further, using cloud computing advantages to remove the need for a dedicated remote location and hardware altogether. Further, the cloud's pay-for-use model reduces monthly charges, prompting organizations to pay for protection only when they need it. Soup-to-nuts cloud-based Disaster Recovery as a Service (DRaaS) offerings, meanwhile, can also reduce the amount of on-staff expertise needed to pull off a DR initiative.
"Before, you needed a large IT staff and partnerships with vendors to make DR work," said Satish Hemachandran, senior director of product management at Cbeyond Inc., a managed services and cloud hosting provider. "With cloud, DR is becoming more accessible."
But while using the public cloud as a platform to recover your systems may finally put DR within reach, cloud DR is far from simple. It represents a different way of doing things, and there are technical challenges to consider. Plus, nascent cloud DRaaS providers vary in terms of the scope, capability and cost of their offerings, and generic public cloud providers have their own security and availability challenges.
Data protection is only half the battle
Though cloud makes disaster recovery accessible to the masses, companies must first understand what a true DR strategy is to take advantage. For some, having a DR plan means having a second copy of their data offsite, whether on tape in a vault or in an Amazon Web Services Simple Storage Service (S3) bucket. But while offsite data protection is certainly an important first step toward disaster recovery, it's not the whole story. For most shops, the goal of a disaster recovery plan is to not only have your data offsite, but to restore and run entire services and applications offsite, preferably sooner rather than later.
Disaster recovery and the cloud
Part 1: Cloud computing advantages bring disaster recovery to the masses
Backing up your servers to the cloud gets you partially to cloud-based recovery, said Doug Hazelman, software evangelist at virtualization backup software vendor Veeam Software. With that, it becomes feasible -- if not necessarily easy or quick – to restore a server in the cloud from a cloud backup.
Many of Veeam's customers use its Veeam Backup and Replication Cloud Edition to make a copy of their VMware and Hyper-V virtual machines in Amazon Web Services (AWS) or Microsoft Azure, thanks to an integrated cloud storage gateway that provides an interface to those cloud providers' object storage services. Where things get tricky is both AWS and Azure run proprietary virtualization environments.
"If you had to do a full recovery, it would be technically difficult because AWS and Azure don't run VMware or Hyper-V," Hazelman said. To get backed-up VMs running there, you would have to convert the backed-up VM from the original VMware or Hyper-V format to the cloud provider's, which may lengthen the process.
If disaster recovery -- not just data protection -- is your goal, a better option is to use a cloud provider that offers the same virtualization platform you have in-house, Hazelman said. Veeam defers to partners such as Verizon Terremark, iLand and Cbeyond to offer disaster recovery using its virtual machine backups on their VMware or Hyper-V-based clouds.
Having your cloud DR site closely match up with your on-premises environment definitely simplifies things, said Anthony Chiaradonna, chief information officer at Consigli Construction in Milford, Mass. About two years ago, the 600-person construction management firm set out to simplify and improve its disaster recovery capabilities, and cloud was the first place he looked.
"We needed a solution that we were comfortable with but that didn't increase our people overhead," Chiaradonna said. "We made a choice not to expand our environment or our staff."
The firm started by ditching tape backups and backing up servers to cloud storage provider SunGard Availability Services instead. A couple months later, when it was confident in the quality of the backup process, it added disaster recovery services for its servers. With the exception of its Exchange environment, which must be available within 15 minutes of an outage, the remainder of Consigli's servers can be recovered from the nightly SunGard backups within 24 to 48 hours directly on the SunGard cloud.