Editor's Note: Recovery Point is a national provider of integrated business continuity solutions for government agencies and a broad array of commercial and non-profit organizations of all sizes.
by Dick Fordham of Recovery Point
With the rapid maturing of cloud-based backup and disaster recovery services enabled by the convergence of virtualization, highly efficient disk-based backup technology and high bandwidth networks, enterprises can now design disaster recovery (DR) strategies driven by their core business objectives -- instead of doing it the other way around.
After all, how many companies can actually afford to be out of business for the several days it could take to retrieve tapes from storage and then restore systems, applications and data if tape provided their only backup?
Today’s disk- and cloud-based backup and recovery solutions support modern day business continuity and DR requirements by storing customer data on multiple mirrored disk systems from which almost any virtual or physical replacement server can be quickly repopulated. In the event of a disaster, companies can resume operations in a matter of hours or even minutes. Emphasizing these advantages, some cloud backup and recovery service providers go so far as to say their solutions eliminate tape.
But does the cloud really mean the end of tape?
No. While no dynamic or data-intensive business can depend solely on tape for acceptable backup and recovery, this year’s highly publicized cloud services failures such as at Google and Amazon demonstrate that neither can one depend solely on the cloud.
Tape can still fill an important role in many companies’ DR plans as part of a hybrid disk-cloud-tape solution; one that assigns different tiers of data to the storage technology that best meets an organization’s varied recovery point and recovery time objectives.
At a minimum, such a solution employs tape for economical and secure long-term retention and archiving of data that does not require the immediate access of higher cost disk storage. After 50 years, time for a better approach
Traditional IT disaster recovery models have been constrained by the limitations of tape backup technology. For decades, companies backed up data to tape on their own premises on a weekly basis, with daily incremental backups to capture changes, and then rotated those copies through off-site storage. In case of disaster, the appropriate tapes would be delivered by truck to a disaster recovery facility, where the backed up data would be copied from them into an exact duplicate of the production data center infrastructure. If a company was lucky, this process could be completed in a few days.
Longer recovery points and times and the cost of duplicate hardware were not the only drawbacks to this traditional approach. The model was strategically upside-down in that a company’s DR efforts could deliver only a speed of recovery that was tactically achievable rather than what was strategically necessary.
Disk backup systems provided a significant advance over tape, enabling more frequent snapshots of ongoing operations and much shorter recovery times. Under such continuous use, disks are also much more reliable than tape. But disk storage systems, in particular high-end storage area networks required by the largest organizations, can be costly. And they are not portable like tape, requiring some other method of replicating backed-up data to an offsite location.
This obstacle was overcome through the combination of several enabling technologies.
First was advanced deduplication and compression that reduced the backup footprint, helping to contain storage hardware requirements. Next was the availability of high bandwidth networks that increased the practicality of transmitting these compressed backup files to remote backup facilities or service providers. Finally, virtualization eliminated the need and cost to exactly replicate production data center hardware at the DR facility.
Companies can now meet their backup and recovery requirements in a variety of physical-to-virtual, virtual-to-virtual, virtual-to-physical and traditional physical-to-physical configurations.
These technologies have now given rise to cloud backup and recovery: the provision of software, platforms and infrastructure as a service to enable an integrated backup and recovery process, accessible from anywhere, at any time, via any carrier. In a complete reversal of the old tape-based model, cloud backup and recovery enables a strategically-driven DR approach in which an organization can set a minimum acceptable time for accessing data and resuming operations following a disaster.
For organizations seeking shorter recovery point and recovery time objectives, cloud backup and recovery fills the gap between the matter-of-minutes recovery delivered by costly, high-end SAN replication and the multi-day recovery achievable with tapes and trucks. Cloud backup and recovery provides same-day recovery, restoring business operations in hours as opposed to days or even weeks, with as little as 15 minutes of data loss all at a lower cost than tape-based processes.
The death of tape is greatly exaggerated
So has the cloud eliminated the need for tape, as some service providers claim? Not at all.
Just ask the 150,000 or more Gmail customers who would have lost data were it not for tape. When Google’s February 2011 update of storage software went awry, it not only emptied users’ accounts but also rendered Google’s redundant cloud-based copies unavailable. Google was, however, able to restore users’ data from offline tape backups.
Unfortunately, many customers of Amazon EC2 cloud services didn’t fare as well when a network change in April 2011 led to a failure that disabled some high-profile sites for days and left many customers with permanent data loss.
Since tape is not mentioned in any media coverage of the failure, one can only assume that the same type of tape backup that saved Google customers’ data was not available at Amazon.
Replication of backed up data to redundant, mirrored physical or virtual storage systems is a disaster recovery best practice, but only if the multiple backups are themselves not vulnerable to the same threat. As unlikely as the simultaneous failure of all backups may be, one important role for tape in the era of the cloud is as the ultimate backup for data that resides, well, who knows where?
A hybrid solution
As demonstrated in the Gmail recovery, secure offline storage is the minimum role that tape should play in a hybrid backup and recovery solution that combines on-site disk backup, physical or virtual cloud-based storage and tape.
Tape can also remain an effective and economical backup choice for platforms hosting an organization’s less-than-mission-critical data. For lower tier data for which longer recovery times are acceptable, tape can be a more affordable solution than disk. A hybrid backup and recovery solution allows IT to devote the appropriate resources to different tiers of data, assuring that high priority data can be made available in the least amount of time, while lower priority data is restored as needed.
Tape is clearly the most economical and possibly the technically superior choice for long-term storage and archiving, especially for organizations subject to any of the estimated 10,000 regulations covering data retention, integrity and security. When tape is used for long-term storage, most problems associated with it become non-issues. Tapes stored in environmentally controlled vaults without being continually handled, mounted, read and written to deliver extremely high reliability.
For the quality and completeness of their own service (if not for the customer’s own peace of mind) a service provider should offer the option of tape backup as additional assurance that in the event of a disaster, the customer will achieve a complete recovery.
The choice and design of a backup and DR approach should always be driven by an organization’s actual recovery requirements: How little data can we afford to lose (recovery point) and how quickly must we be back in business putting that data to work (recovery time).
For some organizations, only the fastest SAN-based replication will suffice. Others will be well served by the latest cloud backup and recovery services.
But in addition to achieving the fastest possible recovery, organizations also need effective redundancy in their backup program to assure that the data required for recovery is available when it’s needed.
As the volume of backed-up data eventually grows, disk storage can become too costly to provide that redundancy. A hybrid solution that includes regular backup to tape can provide customers a cost-effective combination of today’s most advanced backup technologies with the proven longevity of tape stored securely offline.
Dick Fordham is Director of Corporate Strategy for Recovery Point, a leading provider of Disaster Recovery services.