Customers traditionally have looked to the cloud for backup and storage services, but recently they have set their sights on cloud disaster recovery services.
Thanks to some high-profile outages
Failure is not an option, and providers must prove that their cloud disaster recovery services are reliable. Cloud providers can win over customers by building accessible and resilient services backed by strong service-level agreements (SLAs) for data and application protection.
Cloud disaster recovery services: An enterprise-grade option?
Data-replication software vendor Zerto recently announced its Zerto Cloud Disaster Recovery Ecosystem (ZCE), a partnership program aimed at helping cloud providers establish reliable cloud disaster recovery services. Thirty-three cloud providers have already joined ZCE, including Terremark Worldwide Inc., Bluelock LLC and Peak 10 Inc.
The ZCE program allows cloud providers to replicate data at the hypervisor level -- based on Zerto Virtual Replication 2.0 technology -- as opposed to the more traditional way of replicating data through storage architecture, said Gil Levonai, vice president of products at Zerto.
More on cloud disaster recovery services
Managed service provider's guide to selling cloud disaster recovery services
Cloud disaster recovery services shaping up
Business continuity and cloud disaster recovery services for cloud services
Cloud providers offering disaster recovery (DR) traditionally have been required to have the exact storage type as their customers, a requirement that drives up costs for both provider and the end customer. With hypervisor-layer data replication, DR is storage-agnostic, Levonai noted.
"I think a lot of cloud service providers think disaster recovery is a niche offering, but it's a very important one," Levonai said, noting that many cloud providers want to be able to offer DR services, but have been unable to, thanks to limitations with legacy array-based data replication.
"[Zerto and its participating cloud providers] are actually replicating virtual machines, and this is much more scalable, flexible and more in line with the cloud paradigm," Levonai said. "We are replicating data on an ongoing basis, so if lighting strikes, [the provider] can fail over to the other data center and keep the user up and running."
Customers of the ZCE partners are able to select either Disaster Recovery as a Service (DRaaS) for protecting on-premises applications in the cloud, or in-cloud DR for protecting applications deployed in the cloud and replicated to a secondary data center, Levonai said.
Will cloud disaster recovery services help spur adoption?
Customers will need to be completely assured that cloud disaster recovery services can guarantee full recovery of all applications and data in the event of a disaster, assurance that Zerto claims its cloud provider clients can offer their customers.
Enterprises are well aware of highly publicized cloud failures, but the cloud's portability and mobility make it a great application for DR, said Amy DeCarlo, principal analyst for security and data center services at Current Analysis Inc. She predicts that more services like Zerto's ZCE, geared toward cloud disaster recovery, will emerge.
"Providers can use the cloud as a disaster recovery offering for a client that previously may have needed to create an entire replicated disaster recovery site, which would have been very expensive," DeCarlo said. The flexibility of cloud allows workloads to be backed up very quickly and more cheaply for the end customer, she said.
End users can use Infrastructure as a Service, or IaaS, for a variety of applications, and DRaaS is one that is only going to grow in popularity, so long as the services can be guaranteed by the provider, said Sam Barnett, directing analyst for data center and cloud at Campbell, Calif.-based Infonetics Research Inc.
Cloud providers traditionally have struggled with providing the level of replication their customers required, and moving large data sets ate up bandwidth. In developing a viable DR offering for customers, providers must start in the data center, Barnett said.
"The infrastructure is of course the important first step, but then the actual performance must speak for itself," Barnett said. "[Cloud providers] must be able to carry out what they are saying they can do for disaster recovery -- especially with large data sets."
DRaaS needs end-to-end SLAs
Committing to strong SLAs can help providers convince customers of providers' DR guarantees, DeCarlo noted. "There are a lot of questions [providers] are going to have to answer around resilience, and they are going to have to assure stability and the ability to make the data accessible quickly, while not skimping on security," she said.
Unlike SLAs for many other cloud services right now, SLAs for cloud disaster recovery services need to provide end-to-end guarantees. "In a disaster recovery SLA, end customers are going to need to see guarantees that in the event of disaster, the provider can fail over and make data accessible again within a predetermined period of time," DeCarlo said.