LTO-4 on the way: Long live tape

The next generation of the LTO tape format will double the capacity to 800GB and increase the data-transfer rate to 120MBps.

By David Hill

In January, the Linear Tape-Open (LTO) Program announced the availability of licenses for the LTO Ultrium tape format generation 4 (LTO-4) specifications for tape device and media manufacturers. Products based on these specifications are expected later this year (see “Inside LTO-4: Speeds, feeds, and added features,” right).

The first three generations of LTO products have been very successful in the open systems market (see “LTO dominates midrange,” below). A major reason for the success of LTO is the non-proprietary nature of the products. Users get investment protection (as well as competitive price leverage) by having the choice of interchangeable tape drives and media made by a number of competing manufacturers. Another reason for the success is that the movers and shakers of the LTO Program are the three major tape technology providers—Hewlett- Packard, IBM, and Quantum. (Those three vendors are the LTO Program’s “technology providers,” but Tandberg also manufactures LTO tape drives.) The LTO Program stays on track because the three vendors share a common goal that it is in their separate as well as collective interest. Individually, they also offer market and technological expertise that benefits other licensees as well as the tape library manufacturers, resulting in a smooth-running collaborative ecosystem.

But is there trouble in tape paradise? Market observers have predicted the death of tape almost from the time that tape was first used. The latest doom-and-gloom forecast comes as a result of two emerging trends. One is the use of disk-based data-protection technology, such as virtual tape libraries (VTLs), that can serve as a target for the backup and restoration of data instead of tape. The second is the use of data de-duplication technology that can make the use of disk-based data-protection technology much more cost-effective by dramatically reducing the storage space required to store data.

On top of that, each generation of tape technology tends to double the capacity of the previous generation. For example, the LTO Ultrium format generation 4 capacity is 1.6TB compressed, up from 800GB (assuming a 2:1 compression ratio) in generation 3. Won’t users sooner or later need fewer tape drives and cartridges?

We believe these changes will likely impact the tape market from a revenue growth perspective—perhaps staying flat or growing only nominally-over time, but a major decline in tape usage is not in the cards.

The LTO Program only announces a road map of two future generations beyond the current generation, but, if tape were a cat, it might be on only the fourth of nine lives, with at least five generations left to go.

Although both disk and tape capacity/performance continues to expand, these advances have not yet slowed the use of either technology. That trend is likely to continue as long as storage growth continues at the rate it has demonstrated in recent years. But, in the long run, isn’t that likely to change? Probably so, but to quote John Maynard Keynes, the British economist, “In the long run, we are all dead.”

One reason for tape’s longevity is that it’s still much more cost-efficient than disk (even SATA) on a raw per-bit basis, and that is not likely to change any time soon. If a company’s data-protection objective focuses more on data preservation than it does on data availability (as measured in seconds or minutes), then tape is, and will remain, a viable choice.

The need for the high availability of data in an enterprise will continue to grow. However, that is only the visible part of the storage iceberg. Beneath the visible storage “water line” remains the large majority of data. This fixed content data no longer needs to be highly available, but it also consists of multiple—and the operative word is multiple—data-protection copies of online data. At least one data-protection copy needs to be on-site and one copy off-site. And that is the barest minimum.

Although disks can be used for the bare-minimum data-protection copies from an availability perspective, many or even most enterprises will still want to take advantage of tape technologies to create and secure additional copies. This is because having the additional layers of protection represented by additional copies of data is relatively inexpensive compared to the risk of permanent loss. A second reason is that enterprises should not put all their eggs in one technology basket to protect against the possibility of a systemic failure of any one type of technology.

Won’t de-duplication destroy the raw cost-per-bit advantage of tape? Data de-duplication is an exciting technology, but it’s not a game changer with respect to tape. De-duplication is not always an appropriate choice for some types of large unstructured data files, such as medical images.

Enterprises may very well use tape as another storage tier for cost-efficient archiving of data where preservation is the dominant data-protection objective—for example, fixed content data that needs to be kept, but for which access is hopefully never needed. That could include old e-mails that have to be saved for years for e-discovery purposes as part of the changes in the Federal Rules of Civil Procedure, for example, but otherwise have no business use. Or it could be large files, such as medical images, for which recovery in a few minutes would be satisfactory.

Overall, we believe that disk and tape are more complementary than opposing approaches to data storage, and the two technologies are likely to be successfully combined into hybrid solutions (for example, a VTL on disk that works in conjunction with a physical tape library).

New disk-based backup/recovery technologies—such as CDP, MAID, and VTL—are all promising, but tape remains a solid choice in a variety of strategies, including data protection, and for certain kinds of unstructured files. In addition, tape can complement other storage solutions or can be used in a hybrid environment with disk-based technologies.

To paraphrase Mark Twain, “The reports of tape’s death have been greatly exaggerated.” In fact, the LTO Program’s announcement of the LTO-4 specifications is evidence that tape technologies are alive and well.

David Hill is the founder and principal at the Mesabi Group consulting firm. A version of this article originally appeared in the Pund-IT newsletter (www.pund-it.com).

Inside LTO-4: Speeds, feeds, and added features

By Kevin Komiega
New hardware based on the next-generation Linear Tape-Open (LTO) Ultrium format came one step closer to reality earlier this year with the news that licenses for LTO generation 4 (LTO-4) specifications are now available for storage device and media manufacturers.

Spokespeople for the LTO Program announced continued support of their six-generation road map (see figure) by doubling capacity, boosting performance, and adding encryption functionality to the popular midrange tape format.

LTO-4 will feature twice the capacity of the previous-generation 3 (LTO-3) technology, increasing cartridge capacity to 800GB native, or 1.6TB with 2:1 data compression. Transfer rates are also improved: 240MBps for LTO-4, compared to 160MBps for LTO-3. (Both transfer-rate specs assume 2:1 compression.)

Click here to enlarge image

“The native data rate of 120MBps translates to about 864GB per hour,” says Bruce Master, senior program manager for IBM’s Storage Products Division. “This will help users to better manage their backup windows, reduce space consumption in IT centers, enhance library utilization, and reduce the amount of tape handling by people and robotics.”

However, the LTO-4 spec goes beyond “speeds and feeds” by incorporating support for more-advanced data-protection and security features.

Generation 4 will continue to support the write-once, read-many (WORM) functionality that debuted in LTO-3 products and now, for additional data protection, LTO-4 will support 256-bit AES encryption via hardware-based writing of encrypted data to LTO-4 cartridges.

The LTO Program “technology providers” (Hewlett-Packard, IBM, and Quantum) expect LTO-4 products to hit the streets by mid-year. (Tandberg also manufactures LTO tape products.)

The LTO-4 specification provides for drives with backward read/write compatibility with LTO-3 cartridges, and backward read compatibility with LTO-2 cartridges.

For more information on LTO technology, visit www.ultrium.com.

LTO dominates midrange

In September 2006, the LTO Program vendors announced that they had shipped more than 1.5 million LTO Ultrium tape drives and more than 50 million LTO tape cartridges.

Although final figures were not available at press time, LTO-based tape libraries were expected to account for more than 88% of the shipments of midrange libraries last year, up from about an 84% market share the previous year (see figure). That translates into more than 55,000 LTO libraries shipped in 2006 and revenues of about $1.26 billion.

Click here to enlarge image

The LTO format competes primarily with the AIT/SAIT and DLT/SDLT formats. Collectively, these formats used to be referred to as “super drives.”

LTO drive manufacturers include Hewlett-Packard, IBM, Quantum, and Tandberg, while LTO media manufacturers include vendors such as Fujifilm, Imation, Maxell, Sony, and TDK. LTO library/autoloader manufacturers include Breece Hill, Fujitsu, IBM, Overland Storage, Qualstar, Quantum, Spectra Logic, Sun/StorageTek, and Tandberg (which acquired tape drive/library manufacturer Exabyte late last year).

This article was originally published on April 01, 2007