The Past and Future of the Hard Drive

Snags writes "Brian Hayes of American Scientist has written a nice little historical review of hard drive technology, from the first hard drive (nice pic) made by IBM in 1956 to what may be available in 10-15 years. He muses on how to fill up a 120 TB hard drive with text, photos, audio, and video (60,000 hours of DVD's). Kind of ironic that this came in my mailbox today considering IBM's announcement."

Comments

Re:More space..

[heliocentric]

I really don't want to go through the trouble of burning CD's or backing up to tape...There's no reason I should ever have to delete anything ever again..

Ok, I like the idea of having all my stuff handy as well, but you need to think beyond just you deleting something... There are other reasons to backup your data including natural disaster. And no, burning stuff to a single CD and wiping it so you can play back your MP3s isn't a backup - you've just made the natural disaster issue portable.

Uh. Except that I cant find anything I'm looking for anymore.. Can't this search function go faster?

No clue what OS you use, do you have some form of *nix where you've got access to the locate command where it's not actively polling the drive when you request the info (however data can be stale)?

120 TB == one snapshot

[Gis_Sat_Hack]

There are already a number of Terra satellites downlinking data at about 4GB/hr, circling from pole to pole in orbits lasting under 2hrs.

There are multitudes of airborne surveys churning out digital snapshots at 400MB a frame.

Mosaiced together at 1m resolution with R,G,B and mean height above sea level, how much storage will a single global snapshot of the earth take ?

Then consider for historical and environmental reasons, most urban/semi rural areas deserve a mosaiced snap at least once a year.

120 TB is just the start . . .

Re:Reminds me of a conversation I had

[ottffssent]

Lesse here - hard drive capacity is growing at what, 60% per year? The largest consumer hard drive right now is 160G. So, that gives us:
2002: 160G
2003: 256G
2004: 410G
2005: 655G
2006: 1.0T
2007: 1.6T
2008: 2.6T
2009: 4.3T
2010: 6.8T
2011: 11.0T
2012: 17.5T
2013: 28.0T

So, if history is any indication, it will be somewhat more than 10 years (just under 15) before 100TB drives become available at the consumer level.

Despite being able to surmount what were once thought to be intractable magnetic effects limiting hard drive density, I don't believe hard drives will make it past a terabyte or so. We are quickly approaching the point where the energy involved in changing a 1 to a 0 (and yes, I realize that with common data encoding schemes, it's not that simple) is less than the thermal energy present in the system. Just as fickle electrons are being replaced by photons in data transport, I think they will replace electrons in data storage as well. Photons leave each other alone so you can pack them more densely; they're low-power; they're resistant to external forces such as electric and magnetic fields and various forms of radiation.

I'm confident there will be 100TB storage devices commercially available within 15-20 years (to give myself a little wiggle room), but if it's based on spinning disks of any kind, and one of you can find me, I'll eat my shirt.

Re:I'll fill it.

[colmore]

A lot of people are mentioning 3d data but...

There is no way of capturing a fully 3d image

perhaps i'm just low on creativity, but I can't imagine any way of capturing vide from real life with a system that isn't functionally equivalent to some finite number of video cameras.

now lets say that in the future telivision is filmed at 10x DVD resolution and from 10 different angles

that only adds 100x to current video storage needs. nothing to sneeze at, but also not so spectacular that you'd need a petabyte drive either.

once we have the ability to record any sort information at fidelity approaching the maximum for human perception, storage growth will rapidly outpace our need.

people seem to be upset about this though...
once storage reaches the maximum that any user would ever need, it has no place to go but cheaper.

also, file system organization will need to be massively overhauled to make a petabyte drive remotely useful.