Tape storage has been around for decades. It is reliable enough and cheap for backup, archive, and restore, but it has well-known disadvantages: backups can fail; drives can also fail; and restores are slow because of the sequential access involved. With disk-to-disk backup eating into the tape market for enterprises wanting to cut backup time and provide much faster restores, does tape have a long-term future?
Its dominance rests upon it being reliable enough and very much cheaper than disk as backup data mounts up into the terabytes, tens and hundreds of terabytes area. Tape cartridge capacities are heading past 1TB, witness roadmaps from the LTO consortium, Quantum with Super DLT, Sony's SAIT, and both IBM and StorageTek's own tape formats.
The bigger the tape though, the more severe the restoration problem. Just finding the file on a 4TB tape with 8TB of compressed data on it can take several minutes.
Any contender to displace tape would have to offer tape-class capacity, disk-class file restore times, and tape-class media costs. It would also, most probably, need to work with existing backup applications such those from CA, Legato and Veritas.
Not a problem
This is rapidly becoming a solved issue. Tape libraries such as ADIC's PathLight VX v2 represent themselves as tape devices to backup applications, even though the files initially go on to disk. The backend of such libraries are tape drives. Were alternative media to displace tape then they would replace the backend in such libraries, meaning no change to backup applications.
Colossal Storage is run by Michael E. Thomas, its president. Al Shugart, founder of Seagate, the world's largest disk drive company, says,"I don't understand all of Michael's technology but I know this is the way to go for the storage industry." This is not really encouraging.
The company says it can deliver a 10TB capacity 3.5 inch (c 87mm) holographic disk at a cost of $0.08/gb - that's $800. That's encouraging from the tape displacement point of view. A 800GB (compressed) LTO3 cartidge costs under $120. That would be c$1500 for 10TB on tape. This could be very interesting - if it is practical.
Aprilis, InPhase and Optware
Aprilis produces a 120mm - CD-size - 200GB holographic cartridge - 400GB compressed - with a 75MB/sec I/O rate. It positions its holographic products thus: "Aprilis can provide a high capacity, Write Once Read Many (WORM), solution using its proprietary media to bring enterprises a cost-effective alternative to high-speed RAID and Tape back up systems. These systems don't compare to Aprilis in access performance, data density and price." The company specifically mentions medical imaging, document archive and image databases as good applications. There is an element of providing better optical disks here.
We can assume capacities will rise, meaning 400GB native and 800GB compressed. That is nearing terabyte tape catridge levels. Media cost isn't mentioned.
InPhase talks about 200+GB on a 120mm disk, rising to 1.6TB by 2009. The company thinks its technology could replace tape in archive applications, stating, "Generally, new technologies find their niche without completely replacing the older technology. We expect HDS will do the same. In some instances it will replace tape, especially in archival applications that have the need for high capacity and long-lived media. HDS has an archive life of 50 years versus 7-10 years for data and video tape. In addition, HDS has random access and tape does not. In recent years, DVD has been used for archive applications. However, HDS is superior to DVD in capacity and transfer rate."
For InPhase holography basically means a better optical data disk, not a tape cartridge replacement.
Optware thinks its holographic versative disk - HVD, not DVD - could hold a terabyte of data. Its initial offering will be a $100 CD-size disk holding 200GB. The drive will cost $20,000. Given traditional industry developments then we could be looking at a $600-800 disk holding a terabyte.
It looks as if the scanty cost/capacity data indicates that holographic media could cost less than tape cartridges on a per-GB basis and offer far faster restore speeds. That gets them onto the tape replacement stage at least.
But this is only an initial starting point. If the technology delivers working and reliable product and if it reaches terabyte cartridges and if tape can't match its media costs then maybe tape's future is going to be circumscribed.
Watch out for an optical disk library with a hard disk buffer or cache in it that can be written to and read from by existing backup applications. If the industry produces a Blu-ray or HD-DVD version of such a box then the road to a holographic backup storage library is much clearer.
One other point. The existing optical storage companies such as Sony, Plasmon, Panasonic and others are not involved in holography. Nor are the existing tape drive/format suppliers: StorageTek; Quantum; Certance; IBM and HP. Were any of these suppliers to invest in or buy a holographic storage company then that would indicate they are taking holographic storage seriously.