The process of developing a business continuity plan -- and keeping it updated -- for just one company is hard enough. The task of safeguarding dozens or hundreds of customers living on your cloud infrastructure is nothing short of daunting.
By submitting your personal information, you agree that TechTarget and its partners may contact you regarding relevant content, products and special offers.
But by becoming more disciplined about virtual machine (VM) backups and data archiving, cloud providers not only can ensure they will meet the terms of their basic service-level agreements (SLAs) more easily; they also may be able use this newfound resilience as a springboard for creating more profitable, premium offerings for customers seeking a deeper data archiving or disaster recovery service.
"Clients tend not to pay the proper amount of attention to this. Just because you are going to Infrastructure as a Service doesn't mean that you get rid of your backup plan," said Michael Pearson, president of DSA Technologies, an Infrastructure as a Service (IaaS) provider in Elk Grove, Calif.
Bluelock, an IaaS provider in Indianapolis with experience in managing virtual infrastructure, became diligent about VM backups from the company's start, according to Ben Miller, product solutions director at Bluelock.
"We actually started off by treating our own infrastructure as a service that we need to run our business," Miller said. "We back up every VM that is created, no matter how big or small."
Bluelock manages its daily VM backups on a weekly basis, and it retains monthly summaries for up to a year -- although it doesn't necessarily keep all that data within the local infrastructure, moving them to tape archives at a secondary location.
"It's a pretty thorough backup solution for even the most demanding kind of situation," Miller said.
Does your VM backup strategy go far enough to both protect your infrastructure and position your organization to add even greater value for its customers? Here are some best practices that should be on every cloud provider's checklist.
1. Build proper redundancy into infrastructure
The best business continuity strategy starts with ensuring virtual cloud infrastructure includes redundant servers and processors; additionally, the storage area network (SAN) should be replicated across the entire infrastructure, said Jason Pollner, co-CEO of IT Authorities, a managed services provider (MSP) in Tampa, Fla.
More on virtualization in cloud
Pop quiz! Do you know how to control VM sprawl in the cloud?
Check out our OpenFlow tutorial for cloud providers.
Podcast: Seriously, what's so special about Amazon's Reserved Instance Marketplace?
This will help protect cloud providers not only from hardware failures, but also against any human error that could result in a corrupted VM, he said.
"VMs should reside on a storage system that is capable of providing snapshot backups, and they should be stored across redundant hardware," Pollner said. "This covers the short-term backup requirements for data and also provides a complete bare metal restore option."
Meanwhile, virtual disaster recovery vendors like Zerto can help cloud providers offer VM failover to other physical servers in the same data center or to another site within the cloud infrastructure, said Pearson of DSA Technologies. Other tools that help with this task include Dell AppAssure and the VMware vCenter Site Recovery Manager, he added.
2. Figure out what customers need: Bare metal restores or data archiving?
Storing local data snapshots of VMs is a good idea to help accommodate fast recovery of virtual servers and applications that need to be reconfigured or reset, said Terry Buchanan, chief technology officer at Zycom Technology Inc., a cloud service provider and MSP in Kingston, Ontario.
Zycom makes the most recent snapshots readily available and has launched a premium service offering for customers that want fully configured VMs ready to be switched on at a moment's notice.
"Instead of just charging for capacity, we can charge for CPUs [and] for keeping a copy on standby," Buchanan said. "You've got a self-contained instance of a fully working virtual machine."
But just storing VM images locally isn't a sufficient business continuity plan for providers, since the sorts of workloads being hosted by cloud providers are so dissimilar, said Miller.
For example, one customer might be using the cloud as its application-testing sandbox and might not need data restored, while another is running production apps and needs both the VMs recovered as well as the related data for those applications.
That's why Bluelock also works closely with customers to determine the sorts of tasks and applications being handled by its VM infrastructure, matching its deeper data archiving options accordingly.
You have to have an archival strategy on top of the VM system restoration plan.
Co-CEO, IT Authorities
"A client is presented up front with the choices and the ability to roll these into their contract or choose on the fly if they would like to be backed up," Bluelock's Miller said.
Some might choose, for example, simply to invest solely in regular VM snapshots, while others need to supplement those snapshots with regularly archived data -- usually for compliance purposes.
Cloud providers might need to consult with their customers' IT staff to determine which approach is best for them and which might provide other value-added service opportunities.
"[Customers] need someone with business expertise to guide this decision," Pearson said.
3. Build in geographic redundancy
Cloud providers need to think in layers when planning VM backups and recovery processes. While the need for fast recovery means that local access to backup snapshots should be a given for every cloud provider, true redundancy requires that files also be stored at a secondary location.
"You have to have an archival strategy on top of the VM system restoration plan," Pollner said.
In some cases, that means saving VM image snapshots to disk infrastructure in another data center location. It could also mean cloud providers need to transport backup tapes somewhere else for safekeeping, as might be done for a traditional data center.
"We do off-site backups to remove the VMs from the physical location here," Pearson said. "Our tape solution is not intended to offer full data center redundancy, but this absolutely differentiates us from commodity cloud providers, and we introduce it into every sales discussion."
About the author:
Heather Clancy is an award-winning business journalist and an ongoing contributor for several TechTarget sites, including SearchCloudProvider.com.