For cloud backup and disaster recovery, bandwidth proves problematic

When it comes to cloud backup and disaster recovery, organizations are holding back due to insufficient bandwidth and lengthy recovery times.

Small- and medium-sized businesses are increasingly turning to commercial cloud backup and disaster recovery options, according to a recently released Gartner Inc. report, but large enterprises with multiple data centers are leery of cloud-based backup and DR because of insufficient bandwidth and lengthy recovery times.

The open Internet doesn't get you very good performance … latency in cloud backup is a really big issue.

Werner Zurcher,
research director, Gartner Inc.

The report details the potential pitfalls organizations face when relying on cloud-based disaster recovery (DR). Werner Zurcher, a Gartner research director and author of "Cloud Backup Assessment: The Good, the Bad and the Ugly," said the big companies he's spoken to aren't really interested in cloud backup and disaster recovery. Instead, they tend to prefer using their own data centers, mainly due to bandwidth issues and a mix of security and compliance concerns.

As a result, large organizations tend to stick to one of two alternative options: cache cloud backup or traditional backup with cloud storage. Cache cloud backup (sold as hybrid cloud backup by some vendors) involves sending backup data to the cloud but holding on to some data in local caches; Zurcher noted that cache backup works well for endpoints.

Traditional backup involves an appliance that is provided to customers that aggregates data from different backup modes and stores that data in a local cache. According to Zurcher, the key difference between cache backup and traditional backup is that cache doesn't necessarily rely on servers; traditional backup can be described as server-oriented backup.

"[Companies with multiple data centers] come to be more interested in recovering in one of their data centers," he said. "If you have a secondary data center or DR facility, then people aren't as interested in cloud backup; they want to do on-premises backup."

Zurcher pointed to bandwidth issues as the No. 1 culprit many organizations forget about when offloading DR to the cloud.

This seems to be especially true in what he described as "pure cloud-backup scenarios," where all data is both stored and recovered via the cloud. In cases such as remote offices and data centers, a pure cloud play doesn't work because the LAN becomes a huge bottleneck.

"Depending on how slow you're going and which parts you're using and what your Internet infrastructure is, basically because the open Internet doesn't get you very good performance … you end up getting very limited bandwidth over the public Internet through a cloud provider," Zurcher said. "Even if you had up to a 1 Gbps carrier using that connection into the cloud, the distance becomes a mitigating factor, so latency in cloud backup is a really big issue."

Over distances of 1,000 miles between an organization and cloud provider, for example, Zurcher found the average connection speed over the open Internet to be 8.5 Mbps. Such limited speeds means a 50 GB backup and recovery would take half a day; a 100 GB backup and recovery would take 23 hours, which is slower than many businesses will tolerate.

He pointed out that even accelerated Internet services only provided average speeds of 15 Mbps, translating to a 1 TB every five hours.

"That's too slow to make the [cloud] data center workable," Zurcher said. And even if more bandwidth is the key that makes cloud-based backup and recovery viable for an organization, paying an additional amount for an accelerated Internet connection negates one of the main benefits of the cloud: cost.

For pure cloud backup and recovery, Zurcher suggested putting strict limits on the amount of data backed up to the cloud, and looking for cloud providers that are geographically closer, though not so close as to risk the possibility of both the primary and backup sites going down in a disaster. He said provider data centers that are 100 to 200 miles away tend to produce the best performance.

Ultimately, Zurcher feels companies are focusing too much on the speed of the backups the cloud provides and lose sight of making the data recovery process manageable.

"You've got to make sure the recovery works fast as well," he said. "Don't get suckered into looking just at the backup times."

While security and compliance concerns always arise in any discussion regarding the cloud, Zurcher indicated that SMBs largely overcome such issues to achieve the cost-saving benefits.

"[Midsized companies] put the security concerns behind them. They say, 'Hey, it works well, it's secure enough.' For companies that want to rely on cloud backup because it saves them money, they seem to get over the security issues," Zurcher said. "For big organizations, it's still a stumbling block."

Dig deeper on Cloud Data Storage, Encryption and Data Protection Best Practices

Pro+

Features

Enjoy the benefits of Pro+ membership, learn more and join.

0 comments

Oldest 

Forgot Password?

No problem! Submit your e-mail address below. We'll send you an email containing your password.

Your password has been sent to:

SearchSecurity

SearchCloudComputing

SearchAWS

SearchCloudApplications

SearchServerVirtualization

SearchVMware

ComputerWeekly

Close