by Chris Schin

Today, IT administrators face a mind-numbing array of regulations that mandate the protection of data, often including a requirement to house a copy of data offsite. Unfortunately, backing up data using traditional mechanisms is difficult, expensive, error-prone, and highly manual. This article compares traditional backup to a newer replication-based data protection strategy.

Defining “Traditional Backup”
Organizations have used traditional backup systems for decades to protect against data loss. The components of a traditional disk-to-disk-to-tape backup system include:

• Backup software
• Local tape target
• Tape drive(s)
• Tapes (scores/hundreds/more)
• Tape offsiting service provider

Data protection strategies in broad use today evolved in a very different computing era. In fact, today’s four leading backup software vendors (in terms of market share) all opened their code trees in the 1980s, at a time when the first 2.5 inch, 100 MB hard drive had not yet been created, “bandwidth” meant “dialup,” and “broadband” meant “a station wagon full of tapes.”

In the intervening decades, despite the fact that computing infrastructures improved dramatically, the basics of traditional backup as a means to protect data changed only incrementally.

Traditional data protection brings with it very real limitations that could be removed using new technologies. This is done infrequently because these technologies have not been adequately packaged to make adoption easy while still providing a full set of required data protection features.

Newer Alternatives
Today, there are alternative data protection strategies that eliminate the use of backup software and the reliance on tape as part of the backup configuration. These alternative strategies rely on the ubiquity of broadband Ethernet, advances in replication, and file-system technologies like snapshots. They offer faster and more reliable protection than traditional backups, and can unlock the value of the second copy of data in new ways.

One such data protection strategy, sync-and-snap, involves the deployment of intelligent, WAN-optimized replication on the primary data set, directed to a disk-based target housed in a remote data center — optimally one already set up and managed by a trusted service provider —on which is kept a synchronized copy of the primary data set. The second copy is constantly up-to-date, online and mountable, natively offsite, and outfitted with a full, multi-year version history.

Let’s look at leading data protection issues and how they play out with this sync-and-snap alternative:

Backup Complexity
Configuring a disk-to-disk-to-tape configuration (the only way to combine rapid restores with offsite protection in most traditional backup strategies) requires the sourcing, procurement, configuration, and testing of:

• Backup software, typically comprised of multiple application or Operating System-specific agents
• A disk subsystem to function as the primary, on-site backup target
• A tape infrastructure, tapes, and a tape offsiting process (often sourced through third party service providers).

Alternatively, if one deploys a replication-to-service provider approach, all that is required is the installation and configuration of replication software on the primary copy of the data, pointed to the disk sub-system at the service provider data center. This is a more efficient, lower-management approach to data protection that can provide rapid restores as well as offsite data protection.

Backup Windows and Recovery Point Inadequacy
Traditional backup strategies rely on the notion of a backup window due to the invasive nature of backup, as well as the onerous workflow associated with traditional backups, particularly when the backup strategy involves tape and tape offsiting.

Scheduling backups increases the windows of vulnerability for the primary data set — the most recently-modified data is the freshest and therefore typically the most valuable, yet any data that has changed since the last backup (aka the data “tip”) is unprotected. This is a meaningful liability for traditional backup strategies. In sum, backup windows result in Recovery Point inadequacy.

Contrast this with the sync-and-snap approach described above where there is a constant, hands-free replication that keeps your second copy in sync at regular intervals.

Backup Data Verification
Another major challenge associated with traditional backups is data verification. Traditional backup systems modify the data that is captured, converting it to a proprietary backup format that was initially designed to be optimized for tape. This renders the backed up data useless until it has been passed back through the backup software and “restored.” This makes verification of the backups very difficult, since test restores are a luxury that enterprise IT professionals cannot typically afford.

Alternatively, replicating the primary data to another live, mountable data set allows ad- hoc or scheduled data verifications requiring nothing more than a glance at the second copy of the data as it rests on disk.

Version History
Creating a version history with traditional backup software can be difficult since each incremental backup needs to be burned off to tape in order to be well-protected. This makes restoring from versions complicated, because it requires the recall of a specific set of tapes that must be restored in careful order to ensure the re-creation of the particular version of the file that is desired.

Conversely, using a sync-and-snap approach, the file system housing the second copy can be configured with any number of snapshots, which keep fully-instantiated versions of the entire file system and all its files in various points in time, available for direct access and restore.

Restore Complexity and Time-to-Recovery
Since the backed-up data is stored in whatever blob format the backup solution implements, it takes time and cost to get the data back in a usable state. This is particularly true in the event of a pervasive disaster, where the data must be recalled from the tape vault and restored online to a usable state.
If the backup  is created via sync-and-snap technologies and stored online on disk, as a mountable second copy, recovery can be nearly instantaneous — simply remount to the second copy and begin to access the data remotely, while simultaneously replicating it back to your primary site to rebuild the initial copy.

Unlocking the Value of a Second Copy of Data
As described above, traditional backup removes all access to the data that has been protected unless that data is restored back to a usable format. Imagine if that second copy of data was mountable, online, and perpetually accessible using normal data access mechanisms and protocols for read-only purposes. IT administrators could run analysis tools against the second copy to better understand their data set and possibly determine what data would be better stored offline in an archive. They could index that second copy, and use it for eDiscovery. They could use the second copy as a higher-performance copy for access by users who are located near the second copy’s data center. All of a sudden, the creation of a second copy for data protection has morphed into something potentially far more valuable to the organization.

Conclusion
Data protection is vital to any enterprise. Unfortunately, backing up data using traditional mechanisms is difficult, expensive, error-prone, and highly manual. New technologies offer alternatives that can eliminate backup challenges and reduce costs while providing better data protection and faster restores. Stop backing up your data and start protecting it!

Chris Schin is products vice president at Zetta (Sunnyvale, CA).  www.zetta.net