By Salvatore Salamone
April 15, 2004 | Backing up data to disk-based storage systems is becoming an increasingly popular and practical alternative in life science organizations. This backup-to-disk trend, which analysts say is likely to gain momentum, is due to several factors, including the continued drop in price of disk-based storage, the challenge of backing up huge amounts of data in realistic time periods, the introduction of systems that satisfy regulatory requirements, and the increasing demand to keep more data online for longer periods of time.
For some companies, the switch to disk-based backup is for purely business requirements. For instance, the systems biology company Beyond Genomics found that the data it needed to back up had grown in size to a point where its high-performance tape systems could not accomplish the job in an acceptable time period.
‘‘We could no longer meet our backup window of a weekend,’’ says Raul Diaz, director of information technology at Beyond Genomics. Diaz notes that the company has a database of about 9 TB and growing.
About a year ago, Diaz started looking for something to complement the company’s existing backup system. ‘‘We wanted to continue to use our current infrastructure,’’ Diaz says. He chose the Sepaton S2100-ES Virtual Tape Library, a disk-based storage appliance. “To our backup software, it looks exactly like our [tape drive] systems,” he says.
Diaz found that backup sessions that would ordinarily take more than two days now took less than 24 hours using the S2100-ES. (In some cases, it took less than 12 hours.)
Factors Driving Disk-Based Backup
Several concurrent trends are giving IT
managers a reason to migrate from tape to
disk-based backup. They include:
* Continued drop in price of ATA disk drives
* Need to keep more data available online
* Introduction of WORM-like storage software
means data can be written only once, protecting
documents from being altered
* Faster backup, retrieve, and restore times
* Requirements to back up data within a set
And the performance when restoring data was even more pronounced. When data are stored on tape, they are compressed, which makes them more difficult to restore. Another limiting factor is the actual time it takes to get the data off the tape. With the Sepaton system, the disk-to-disk data transfer is done over a speedy fiberlink. With the S2100-ES, Diaz found that data restores were four to five times faster than before.
Diaz plans to use a combination of the S2100-ES and his existing tape drives. The S2100-ES will provide a level of real-time backup and recovery; the tape systems will remain in use for long-term data protection and disaster recovery.
Compliance Is the Key
For many companies, tape and optical storage (i.e., CD or DVD) backup systems still have several factors in their favor. First, they are often less expensive than disk-based systems with comparable capacity. But this is becoming less of an issue.
“Cost [of disk storage] has dropped dramatically,” says Pushan Rinnen, an analyst at the Gartner Group consultancy. Rinnen notes that disk storage tends to cost more than tape, but IT managers today are often balancing modest price differences against the other benefits of using disk systems.
With tape systems, “you search data sequentially, and robotic arms that fetch cartridges all take time to retrieve data,” Rinnen says. “Disk systems give you faster response [due to] their ability to do random searches” for data.
Another important quality of tape and optical storage systems is that they are like WORM devices Ñ write once, read many Ñ a characteristic that can be very important for regulatory compliance. Once data are archived on a WORM device, the information cannot be altered or deleted.
This is an important factor in meeting both HIPAA and 21 CFR Part 11 requirements.
Last year, a number of storage vendors, most notably EMC and Network Appliance, introduced regulatory-compliant, disk-based storage systems that ensure data cannot be tampered with once they are created.
And within the past few months, IBM, Permabit, and other vendors have introduced storage systems with WORM-like features to help organizations comply with regulatory requirements.
Life science companies are finding that these WORM-like features give them an alternative to traditional storage approaches. That was the case with the Bronson Healthcare Group, which had been using an optical jukebox to store radiological imaging data. The system was adding about 1.5 GB of data per day (about 3 TB of data have been stored over the past two years). And last year “we started looking for a cost-effective solution based on ATA spinning disks,” says Scott Dent, a systems administrator at Bronson.
Bronson Healthcare was already using NetApp NearStore file servers, so it chose a NetApp software package called SnapLock Compliance, which allows a portion of a storage disk to be designated a WORM drive or volume.
Once data are saved to this disk space, they cannot be modified or deleted, and an administrator can set a retention date for each file. “The SnapLock WORM [technology] ensures data integrity as the files are put into the system,” Dent says.
Gartner’s Rinnen notes that such use of disk-based systems with WORM-like features is likely to grow. “These systems go beyond compliance; they deal with data-retention management and address how companies archive data for the long term,” Rinnen says.