Slashdot Mirror
Turner Testing Holographic Storage
Izmunuti writes "An article in ComputerWorld describes tests by Turner Entertainment of a holographic storage system from InPhase Technologies as a possible replacement for magnetic tape for storing their movies and other programs for playback and broadcast. The article states that each holographic disk holds 300 GBytes." Even more impressive is the cost per terabyte estimated for just a few years down the road.
Can't find anything about shelf life or connection types. All I could find was that the data was stored in parallel at a million bits at a time.
Also, 27MB/sec, could that be a typo? Seems awfully slow, no?
Mind you, this is hardly a unique problem, only a large-scale concentration of a wide-spread one.
The storage solutions are much more lacking in speed/reaction time than in size.
What I would like to see is not a 1TB harddrive, the size I can get today by buying two harddrives, but rather:
Speed: It is a real bottleneck, to wait for disk access. SCSI is expensive for the home user still.
Throughput: What, still under GB/s ?
Reliability: Since a harddrive is capable storing more and more data, it is more and more important to increase reliability, It takes time to fill up a hard drive, it takes a lot of effort if its a lot of data to backup, so more reliable hard drives would eliminate a lot of problems. I don't care about guarantee, that they exchange the disk if it blows up in x years, my data is still lost then. Let's not even talk about what happens if it's over guarantee period. I'd expect a hard drive to work for five years or so flawlessly, more isn't needed since the technology gets obsolete in that timeframe already.
It takes a man to suffer ignorance and smile
Be yourself no matter what they say
Because the technology is optical, it shouldn't really be thought of as an equivalent of CD or DVD, but more akin to tape technology for long-term backup. In that respect, its current lab throughput of 27MB/sec is comparable to LTO2 tapes, the projected 160MB/sec for the production version is much higher than most backup technologies today.
While there is a market for big and fast storage, there are ultimately trade-offs between the two. 1TB in a 12cm disk is going to have some physical limitations, firstly the maximum speed of rotation before the disk breaks itself apart and secondly the data density which if it's using holographic methods will be using the volume of the medium, not just the surface. taking those limitations into account, it's clear to see that improvements in access times and transfer speeds lie with the accuracy of the seek and read hardware, which will be improved over time based on past optical technologies.
So I'd see this initially as a volume storage backup solution primarily, not a hard drive replacement, or even a home-user class backup system. The industry that I work in would be a prime user of this kind of storage. for instance, we're shooting a 28 part stop-frame animated series in High Definition, current calculated storage requirements for all those frames is 14TB. Online storage for editing is a bank of three Apple Xraids with around 8TB in each one, but how do we archive it? Currently it's dumped onto HD tapes, but eventually we'll need to keep the raw stop-frame footage, and we'll need to find 14TB of space somewhere. In that event, 14 holographic discs totalling $1400 is much more appealing than archiving to 1555 dual layer DVDs or 70 tapes, regardless of the amount of time taken to access it, which will most likely be to restore it to the Xraid for future editing and reference.
I don't understand why companies like this don't opt for just sticking with redundant online storage as opposed to offline such as this. With online, as you upgrade over the years, your archived data gets moved along with, and thus you've no worries about obsolescence of your media or reader. I've heard the argument that the storage space is too costly, though that wouldn't seem to pan out. As time goes by, the MB/GB/TB per dollar will increase, and that data you have archived will become trivial in size pretty quickly. I would think, at least in this sort of application, that a good SAN (where storage is essentially abstracted) would negate the need for this.
If it works that's some pretty impressive technology but I suspect it has a few problems that aren't mentioned that are currently impossible to solve. What makes me think this is the way that all the major electronic manufacturers aren't falling over themselves to buy this company or developer their own version. If this really worked the first person to market would make a fortune. Who knows, maybe they have soved the difficult problems. It would be good if they had.
I used to have a better sig but it broke.
The number you get is not 177 Mbs per second, it is 177 MBps (Mega Bytes per second as you say), whyle the number from CNet is 19.25 Mbps (mega bits per second), and it refers to MPEG2 compression of the stream.
In this case there is plenty of bandwidth available because 19.25 Mbps = 2.40 MBps < 27 MBps. In any case, MPEG4 compression can do much lower bitrates for HDTV video, but I strongly doubt that they will use any kind of lossy compression for their stored archive video.
Why erase? Write-once is fine if you can get the cost down. You can always just write on top of the existing diffraction patterns in order to make old data unrecoverable.
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"