Every enterprise needs archival storage to meet compliance requirements and address litigation issues, but "deep" archiving remains a challenge. Nobody wants to keep disks powered, spinning and serviced for up to 50 years or more. Tape is removable and securable, but tape carries its own long-term readability and reliability concerns. Optical storage is emerging as an attempt to fill this gap, and holographic storage may emerge as the next vehicle for long-term offline archival storage, bringing a mix of large capacity and decades of media stability.
However, holographic storage technology is far from being the next big thing. It has been on the drawing boards for years, and even though most of its technological components are well-founded in current CD/DVD devices, practical holographic storage systems are still in development. In fact, there are really only two principal vendors. This article examines holographic storage technology, highlights its anticipated deployment and considers the potentially rocky road ahead for this high-capacity optical storage scheme.
What is holographic storage?
Holographic storage works by storing a sequence of discrete data snapshots within the thickness of the media. The storage process starts when a laser beam is split into two signals. One beam is used as a reference signal. Another beam, called the data-carrying beam, is passed through a device called a spatial light modulator (SLM) which acts as a fine shutter system, passing and blocking light at points corresponding to ones and zeroes. . The reference beam is then reflected to impinge on the data-carrying beam within the media. This creates a three-dimensional refraction pattern (the "hologram") that is captured in the media. Holographic storage uses circular media similar to a blank CD or DVD that spins to accept data along a continuous spiral data path. Once the media is written, data is read back using the reference beam to illuminate the refraction.
This three-dimensional aspect of data recording is an important difference between holographic storage and conventional CD/DVD recording. Traditional optical media uses a single laser beam to write data in two dimensions along a continuous spiral data path. In contrast, prototype holographic storage products save 1 million pixels at a time in discrete snapshots, also called pages, which form microscopic cones through the thickness of the light-sensitive media. Today's holographic media can store over 4.4 million individual pages on a disc.
Today, holographic storage is a WORM technology that relies on light-sensitive media housed in removable protective cartridges. Although rewritable media and drives will appear in the next few years, much like the progression from CD-R to CD-RW or from DVD-R to DVD-RW, experts note that the most likely application for WORM media is for long-term archiving.
What are the benefits and drawbacks of holographic storage?
The argument in favor of holographic storage is quite limited at the moment, and the value proposition is challenging at best. On the plus side, long-term media stability and reliability is a compelling advantage for deep archiving purposes -- disks and tape simply cannot assure reliability out to 50 years. "Disks are very impervious to the elements," says Brian Garrett, at Enterprise Strategy Group (ESG). "I've seen demonstrations where they dip the platters into something boiling and freeze them and roll them around in the mud, clean them up and they're still usable."
Holographic technology also provides portability, allowing the distribution of dense data that cannot be sent conveniently over networks, such as broadcast or high-definition video. The technology should also become more appealing for shorter term backups and archives as companies continue to rely less on tape backups. For example, holographic storage attached to a virtual tape library (VTL) system might be an excellent tape replacement.
On the downside, early holographic storage drives will run in the $20,000 range, with media costing about $200 per disk. Holographic media capacity is also limited to about 300 GB. While this capacity is expected to grow substantially over time, it's hard to make a case for a 300 GB optical disk against readily available 1 terabyte (TB) hard drives or 1.6 TB (compressed) LTO-4 tapes without a specific application. Furthermore, the long-term reliability and readability of holographic drives is still unproven.
Holographic recording is also very data sensitive. "With holographic, you have to keep the data streaming," says Greg Schulz, founder and senior analyst at the StorageI/O Group, noting that it's not yet appropriate for partial recordings. This is similar to early CD-R or DVD-R systems that required constant data in the drive's write buffer. If the buffer emptied during a write process, the CD-R or DVD-R recording would fail and the disk would be ruined. It wasn't until much later in the technology's lifecycle that "multi-session" and "burn-proof" techniques were added.
Lesser-known drawbacks to holographic storage include light sensitivity and limited shelf life of unexposed (unrecorded) holographic media. Blank optical CD/DVD media is forgiving in its handling and unrecorded shelf life. On the other hand, blank (unrecorded) holographic media behaves more like unexposed photographic paper. Prematurely exposing the holographic disks to light can expose and ruin them, and the unexposed media only has a shelf life of about three years.
Standards are also a concern. The European Computer Manufacturers Association (ECMA) has published two standards in mid-2007 to address Holographic Versatile Disc (HVD) products, dubbed ECMA-377 and ECMA-378. But holographic storage in general has no substantial standards endorsed by the International Standards Organization (ISO). This lack of standardization can work against holographic storage by complicating interoperability between media and drives.
How are holographic drives specified and deployed?
Ultimately, any discussion of holographic storage deployment is theoretical because there are no commercial products available today. Beta products are being evaluated, but manufacturers, like InPhase Technologies Inc., are keeping their beta users under wrap. Consequently, there is no word from the field about value, performance, reliability or any application of holographic products. Still, there are important trends worth noting.
As with most storage devices, the key issues to consider are capacity and data transfer rates. Although holographic storage capacity and performance are currently below current disk and tape systems, they compare favorably to existing optical storage devices. Today, holographic storage media holds 300 GB (uncompressed), and beta drives from vendors, like InPhase, are expected to utilize that media. The InPhase product roadmap touts uncompressed capacities up to 1.6 TB over the next few years. Holographic drives, such as the fledgling InPhase Tapestry 300r, cite data rates of 160 Mbps. Seek time can be a lengthy 250 milliseconds, and you can expect almost two seconds to load or unload the disk cartridge.
The SLM is a critical part of the overall drive capacity and performance. SLMs in today's early drives use a 1,000 x 1,000 pixel matrix (1 million bits) to modulate laser light and encode each data page. In order to increase storage capacity, SLMs must eventually become finer (offering more bits) and switch faster. This will fit more and larger data pages on each disk and allow the drive to write and read more data per second.
Early generation holographic drives appear positioned as single-disk external products connected to the local area network (LAN) or storage area network (SAN). As an example, the Tapestry 300r is expected to provide SCSI, 4 Gbps Fibre Channel optical, Gigabit Ethernet, SAS, and iSCSI Ethernet connectivity options, allowing the drive to reside on a wide range of LAN/SAN architectures. When used with a server, holographic storage devices will invariably require device drivers that correspond to the operating system in use.
Optical storage technologies use lasers for noncontact read/write operations, and holographic drives should also be maintenance free. This is a substantial advantage over tape drives, which require frequent cleaning to remove accumulations of magnetic particles from the read/write heads.
What is the future of holographic storage technology?
The future of holographic storage is fraught with unknowns. "This technology is very promising. I've been hearing about it for years," Garrett says. "But at this point, the No. 1 concerns are [high] cost and [product] immaturity." Experts agree that capacity and performance will only increase over time, moving from 300 GB to 800 GB and finally on to 1.6 TB over the next 48 months or so. But the pace of improvements will ultimately rest heavily on industry acceptance. Given that holographic technology is currently geared toward a niche in the storage market, it may be years before early product releases give way to more capable and cost-effective systems that appeal to a larger storage audience.
Experts also note the possible introduction of "hybrid" holographic media. Just as magnetic hard drives are starting to incorporate significant quantities of flash or RAM within the disk, near-term holographic storage media may add some amount of flash memory in the cartridge to provide a degree of rewritability until a suitable rewritable media is developed and productized.
Backward compatibility also remains a significant unknown. No tape drive in your enterprise today is capable of reading a tape written 50 years ago, and the same specter is in the cards for holographic storage. For example, the InPhase product roadmap suggests a third generation of holographic drives in roughly four years and promises backward compatibility with the previous two generations. (Back to first generation in this case.) Well, then what? If the fifth or sixth or 10th generation drives cannot read the holographic disks written today, you'll need to either retain the older drive software and hardware, assuming that it still functions, or rewrite the older disks to the newer media later -- defeating the purpose of such long retention. "The same corner case that justifies holographic storage also works against it," Schulz says.