Disk-to-disk backup: pass the data please. (Tape/Disk/Optical Storage).

From the earliest days of electronic data, backup storage media was a "whatever performance at the lowest cost" choice, a theory that historically pointed to tapebased backup. But now the evolution of technology is changing the relationship between cost and performance and, much like punch cards and paper tape, tape-based backup carries a lot of hidden costs and hassles that make it increasingly less than ideal. For this reason, end-users are increasingly looking for alternatives to traditional backup-and-restore methods, and disk-to-disk backup is looking more and more promising as a way to eliminate backup hassles, improve backup performance, and reduce overall costs.

According to the Gartner Group, backup remains the single most costly storage administration task. Managing a tape backup system requires complex configuration, continuous monitoring and frequent tuning. Tape libraries need to be loaded with media, tapes need to be manually tracked and moved around, tape drives frequently fail and tape libraries need constant maintenance and attention. Although tape drive capacity and performance are improving, the improvement is not keeping pace with significant advances in disk technologies or the increasing amount of data being stored on a typical network. These problems are compounded because as the amount of data grows, tape devices and their I/O performance do not scale appropriately. Couple this with a lack of skilled IT resources and the ever-increasing demand for high data availability, and the result is that backups place an immense burden on IT departments.

Enterprise backup and recovery means more than disaster recovery to today's enterprise. The rapid restoration of damaged files or file systems (including database files) is critical to maintaining routine business operations. The majority of real-life recoveries in our current IT environment are not "full system disaster recovery" restores; instead, they are partial file system or even single-file restores required to quickly bring critical applications back online and preserve business as usual.

Cost Matters...Again

Historically, disk technology has been orders of magnitude more expensive than tape media. Tape has therefore been used for data intensive applications such as backup and archival. Times have changed. Over the last ten years, the cost of disk has plummeted; it is highly cost-effective now, and its cost advantages over tape will continue to grow rapidly. When it comes to the basic cost of media, tape has lost its edge.

Without even taking the cost of a tape drive into account, it is interesting to compare the costs of disk versus tape media. Only ten years ago, disk cost $1,000 per GB and tape media cost $10 per GB. The recent announcement of a 300GB 3.5inch disk available this year will, for the first time, mean that the per-gigabyte cost of disk will be less than that of tape media alone. More significantly, it shows that in only ten years the cost of disk has declined over 1 000X compared to tape media, which has only declined by l0X. For some time it has been possible to buy a powerful, functional disk array for much less than the cost of an equivalent tape library.

ATA drives today cost about 1 to 1.5 cents per megabyte, compared to 3 to 5 cents for SCSI and 6 to 8 cents for Fibre Channel drives. Tape is typically a penny or less per megabyte, but the cost of media to keep up with storage demands can and does inflate the overall expense. New, higher capacity ATA drives are imminent and anticipated to bring ATA drive costs down to one-half cent per megabyte. Additionally, today's ATA RAID leaders are delivering subsystems that bring the cost of ATA RAID-based disk storage below $3 a gigabyte.

So Much Data, So Little Time

Beyond the changing cost-percapacity of storage media, disk storage can reduce the operational and administrative costs of backup storage. The slow seek time, throughput performance, media characteristics and sequential access nature of tapes and tape drives causes complexity, unreliability and limitations within existing backup systems. Compared to tape, disk offers enormous benefits in terms of performance, reliability and ease of administration.

Slow Tape Seek Time: Optical media and slow-loading tape drives, the solution typically chosen by most nearline storage users, provide access times that range from 10 seconds to more than 20 seconds, including 5 to 10 seconds for a robotic media exchange. In contrast, disk systems provide nearly instantaneous access, measured in just milliseconds.

Throughput and System Performance: The fastest tape drives today perform at transfer rates of 30MBytes/sec. Even desktop class ATA disk drives achieve 40-50MB/sec transfer rates. And the newest generation of enterprise class Serial ATA drives is delivering upwards of 60-75 MB/sec performance-and doing it at prices still one-half that of SCSI drives. In addition, restoration is significantly faster because access to data takes milliseconds compared with minutes in a tape library.

Media Characteristics and Functionality: Ironically, although the vast majority of current information protection is performed by backing up data onto some type of magnetic media, tape backup and restoration often does not work and creates a huge headache for most businesses. Typically, IT departments spend between 20 and 30 percent of their manpower dealing with tape backup problems. Over 30 percent of all storage problems stem from tape backup systems. A recent study found that while tape backups are used extensively, restoring data from a tape backup system fails an astounding 70 percent of the time. The reasons for such an alarming rate of failure range significantly--and may vary from bad tapes or tape drives to the inability to find the backup tapes or careless processing by IT staff. Tape drives and physical tape media can be unreliable and frequently cause backup failure. Conversely, disk-based backup makes use of highly reliable Serial ATA disks coupled with RAID hardware that provides multiple layer s of data protection combined with extremely high MTBF numbers for strong data protection. The maintenance of the disk-based backup solution is far simpler than a tape library, as there are no robotic arms or tape drives. Consequently, maintenance costs for disk-based solutions are far lower than for equivalent tape libraries.

Sequential Versus Random Access and Faster Data Recovery Smooth recovery from operational or data disasters is a competitive necessity and time-to-recovery directly affects an organization's bottom line. With the slow performance of tape-based backup, restoring operations can take many hours even under optimal circumstances. While most failures only involve a limited number of files, the entire backup set must be restored from tape to disk, regardless of whether a full restore is needed or not. Since data on tape cannot be individual data can only be restored, copied and used after a complete, lengthy extraction.

On the other hand, the high throughput and instant access provided by disk-based backup results in much faster data recovery than from tape. Disk-to-disk backup arrays can also restore an individual file or set of data in seconds or minutes without the slow triple step "restore-copy-delete" process. By directly accessing and restoring individual files, it's possible to restart systems in minutes using disk-based backup, dramatically reducing the business impact of emergencies or other unplanned system failures.

Key Features of ATA RAID-Based Disk-to-Disk Backup

Accelerates Data Backup: Backing up primary storage to disk-based backup storage is two to four times faster than tape, shrinking the backup window down to a few hours or less.

Speeds Data Recovery: Administrators can rapidly restore backed up business-critical data much faster than from tape, almost instantly restoring individual files and folders with no tape load time.

Improves Backup Process: Time-consuming tape backups are offloaded from primary storage; backup of several systems is automatically consolidated.

Provides Online Access to Archived Data: Reference and archive data can be quickly retrieved and maintained on cost-effective storage.

Where Did All of This Data Come From?

In today's storage environments, applications being used and data access needs significantly affect storage media decisions. For applications where readily available data would be beneficial, using disk for nearline backup storage provides added application benefits. One such scenario is the overflow of reference data, actually a driving force that favors ATA disk-based storage. According to research by the Enterprise Storage Group (ESG), reference information--such as legal documents, e-mail, presentations, source code, Web content, CAD/CAM designs, medical images, photographs, geophysical data and historical documents--represented 37 percent of all stored data in 2001. By the end of 2005, ESG predicts reference information will represent 54 percent of all new information, a significant amount of information and likely to benefit from nearline availability.

Similarly, many IT operations use tape-based solutions for long-term archiving. Disk-based backup is typically not considered a feasible solution for records that must be archived for years or on a permanent basis. Even in an environment where records are being archived on tape, many IT departments will use disk-based storage as a staging server, recording multiple backup cycles before being archived on tape. In this implementation, while tape may prove useful in maintaining archived records, actual restorations are typically made from data that was saved in the first two or three backup cycles, providing another clear justification for a disk-based solution in the system.

Pass the Data. Hold the Check

Disk-to-disk backup is now a viable enhancement, or alternative, to disk-to-tape operations--due to the emergence of ATA-based disk arrays that are lower in price than tape libraries, as well as the performance advantages of disk drives over tape. Tape remains a strong contender for long term data archiving and off-site storage. System administrators who use disk-to-disk backup can expect a 2X to 3X performance improvement in backup-and-restore operations compared to traditional tape systems. The performance improvements of disk-to-disk operations are directly related to the dramatically faster speeds of ATA drives. And since disk drives are random access devices, data can be written and read (backed up and restored) on a file-by-file level or on whole data sets far faster than tape. Restores can be completed in seconds rather than hours, and the restore risks and dependencies associated with tape are eliminated with a disk-to-disk approach. After data is backed up to disk, further protection steps can be tak en, including cloning or data staging from the disk to tape.

Declining disk drive costs and improved speed are key motivations for end-user adoption of disk-to-disk backup. Related benefits include improved end-user response times (due to faster restore rates) and decreased demand on IT staff resources because disk arrays do not require operator intervention. Improved reliability of disk drives--and the fact that disk-to-disk backup makes it easier to do more frequent backups which ultimately increases data protection--are also critical factors allowing more organizations to improve their backups and restores. And that is just plain smart business. Pass the data please.

www.promise.com

[GRAPH OMITTED]

Andy Newbom is product marketing manager, Advanced Storage Group, at Promise Technology, Inc. (Milpitas, Calif.)