When will we be rid of Hard Disks?

[aak97]

Hard disks are getting bigger and faster, but still, it spins and crashes if you accidently bump into it. When will we be using just Ram or other Non Mechicial devices as secondary storage? 200+Gb USB jump drives (very unlikely)? Or those 3D optical storage I heard about 5yrs ago?

any "Active" research going on?

 

[Peter]

Simple: Cost. Want a 100 GByte FlashROM drive? No problem. Here ya go ...

You're going to spend a little more than on a rotating HDD, though. And they're also a little slower.

 

[buleyb]

Hard drives won't be around forever, but they will last awhile longer. No matter what tech comes around, it won't be cheap to buy (regardless of how cheap it is to make). And unless the industry as a whole gets it's act together and begins embracing single or few tech solutions (the DVD+/-RW fiasco, for example), the market will be a mess.

 

[Mday]

some companies are replacing tape archival with HDD archival.

 

[gherald]

IBM is doing sove very active research, that is part of why they offloaded their disk business to Hitachi.

HDD archival is definately the way to go, especially if the main backup is on a raid 1 or 5 and gets a full 7 days worth of backups mirrored weekly to a semi-remote JBOD second backup. Gigabit comes in very handy, of course.

 

[gherald]

buleyb's sig
When in doubt, kick it and run like hell...

I keep parsing that as: "When in doubt, klick it and run like hell..." :Q

 

[dkozloski]

An interesting concept I saw was to use media that works optically the same as a DVD or CD-ROM but instead of a disk it's a long strip rolled up like scotch tape. It's written to and read from by interferring laser beams that penetrate the layers of the roll. The only limitation to the number of layers is the optical clarity of the tape. This process gives you the physical storage density of magnetic tape but it can be accessed randomly rather than having to be unrolled and read serially like magnetic tape. Compare the surface area of a strip of tape 2 inches wide and 250 feet long with a DVD disk and the amount of data that could be written to one roll is staggering. I think the Norwegians are working on this.

 

[buleyb]

sounds cool, but I'm not holding my breath on that one...

 

[aak97]

the problem I've with HD is they are getting TOO BIG. It'll be really hard for a home computer user to back up 100Gb HD (let alone those 320Gb ones). I'm stupid enough to have a 120Gb Raid0 setup on my old computer.. now one of the 60Gb died and I'm totally screwed.

Another problem w/ HD is kinda like CPU (die shrinks) nowadays. they're packing more data per square inch on the disc, but how far can they go? there has to be a physicial limit to how much data can fit on a square inch right?

 

[Mday]

hmm, i meant that they were investigating replacing long term tape storage with long term HDD storage.

 

[gherald]

Originally posted by: aak97
the problem I've with HD is they are getting TOO BIG. It'll be really hard for a home computer user to back up 100Gb HD (let alone those 320Gb ones).

Home users can get smaller drives, then. It's not like they are going away. 80gb is about standard nowdays, though 120s and 160s are gaining popularity as more and more people discover P2P.

I'm stupid enough to have a 120Gb Raid0 setup on my old computer.. now one of the 60Gb died and I'm totally screwed.

There's nothing stupid about a R0. Stupid is not backing up. BTW "120gb Raid0" sort of implies 2x120gb drives. If you have 2x60gb drives, I would say "60gb Raid0". Because, a Raid0 of total size 120gb could be 3x40, or 4x30, or 6x... you get the idea.

Another problem w/ HD is kinda like CPU (die shrinks) nowadays. they're packing more data per square inch on the disc, but how far can they go? there has to be a physicial limit to how much data can fit on a square inch right?

How about DNA? We're not near that level yet, and I am sure there is plenty room at even more microscopic levels... just you wait ~10 years.

 

[MadAd]

Originally posted by: Peter
Simple: Cost. Want a 100 GByte FlashROM drive? No problem. Here ya go ...

You're going to spend a little more than on a rotating HDD, though. And they're also a little slower.

Any idea how much a *little* more is? I cant seem to find a price for the ata ffd's anywhere.

 

[lexxmac]

Originally posted by: MadAd

Originally posted by: Peter
Simple: Cost. Want a 100 GByte FlashROM drive? No problem. Here ya go ...

You're going to spend a little more than on a rotating HDD, though. And they're also a little slower.

Any idea how much a *little* more is? I cant seem to find a price for the ata ffd's anywhere.

I'd say it's a safe bet that 'a little' means a freakin' lot.

DNA data storage sounds like a great idea to me, but I'm curious how you would read the data, much less write to it. If anyone knows how that works, I'd like to know.

 

[Mark R]

Originally posted by: MadAd

Originally posted by: Peter
Simple: Cost. Want a 100 GByte FlashROM drive? No problem. Here ya go ...

You're going to spend a little more than on a rotating HDD, though. And they're also a little slower.

Any idea how much a *little* more is? I cant seem to find a price for the ata ffd's anywhere.

The 90 GB version is just over $27k

 

[Mik3y]

the future of harddrives will actually just be a small glass cube with 4 lasers beamed right into it. for example, they are just like glass cubes where lasers are marked inside of it to create a 3d picture or wutever. one glass cube is said to be able to hold tens of terabytes of data.

 

[Farmer]

Mik3y:

Interesting, glass cubes with lasers through them, that creates pictures, holding terabytes of data. Sounds almost laughable (no offense). Can you clarify?

If this was the 5-year old optical technology mentioned earlier, I must've missed the news.

Heh, those FFDs are probably insanely priced. "MIL-STD 810F" compliant = $$$.

 

[smahoney]

The biggest problem I see isn't how large disk drives (or any other theoretical storage system) are getting, it's that the pipe for getting information off/out of them is not keeping pace. A multi-Terabyte hard drive sounds nice, but how the heck is anyone going to back it up. A new 250GB drive holds 200 times as much data as an old 540MB drive but the data transfer rate isn't 200 times faster and the access rate is not 200 times less. Memory bandwidths for CPU have scaled better than this. This is an ongoing problem for storage engineers such as myself in looking at the size and scaling of enterprise storage systems. I would be surprised if you don't see some type of inexpensive asynchronous mirroring software for home backups available soon - otherwise it will just take too long to back everything up. Snapshots are nice too but you have to be able to store them on a different spindle. I don't even want to contemplate restoring data to a 4TB disk drive in a few years...

 

[warhorse]

grr, how do I delete this thing

 

[warhorse]

(Sorry, haven't been here in awhile and I messed up the log in process.. and I can't figure out how to delete my last post.)

I sat in on a Seagate talk over the summer, and magnetic hard disks are not going away. I can't remember exact densities, but I do remember some of the techology that they're working on.

First consider the current disk layout. The head flies over the platter and looks for magnetic fluctuations that represent data. For example the head would read a + then the platter rotates a bit and then it reads a - and interprets that.

Once they can't shrink the distance between the +/-0 areas (there's a simple key term for that I'm forgetting) they're going to be putting the fluctuations perpendicular to the platter which means that I think they have to read the top and bottom of the platter at the same place which I think would increase densities but also half the number of surfaces. It just occurred to me that I'm not sure how they're going to be able to position the arm so that it lines up on the top and the bottom. Perhaps seperate actuators for the top and bottom of the platter?

After that when densities get really high, they're going to start using lasers. As I understand it, they're going to heat the platter right by the head with the laser so that the platter expands slightly. This allows the head to more easily read the data as the laser is more accurate than the head but the laser can't read magnetic data. With the area expanded, the head can then read the desired data.

So yes, they've still got several orders of magnitude to go. I'm not too familiar with the ATA bus though; I don't know if they'll run out of bits to represent sectors by then. Hopefully we'll be past 512 bytes per sector before too long.

 

[zetter]

I don't think that Hard Disks will be going away any time soon, especially when companies like Seagate are currently working on putting a Terabyte on a harddrive platter and other techniques like warhorse mentioned in the previous post.

 

[xts3]

You'd think they'd do something to make the head's actuators and the motors more reliable. Seriously they could grow sales on marketing the drives reliability alone I'm sure. If a cars motor went in first 3 years of its life you'd have a lot of mad car buyers.

Failure rates for most consumer drives that see regular use are horrible. Most fail within first 2 years. Out of all the WD drives I've owned none have lasted the beyond 3 years and I've owned over 10. The HD industry has a 'designed to fail' policy for consumer drives. Witness the IBM fiasco. More money should be spent on reliability, and making them more homegrown-repair friendly then upping the density IMHO. I have really old drives that are still running today 120MB-540MB. All my really old drives have outlived the newer drives by a ratio of at least 2:1. I'm thinking its too profitable and cheap to make drives that fail me thinks. If HD companies would sell a decent sized drive (>120GB) with 5 year warranty I'd be all over it. WD is doing good with the Raptor series of drives but the price is a little high for the size of the drive itself, even if it is 10K RPM, I'd like to see a drive with 7200RPM /w 5 year.

Any industry insiders want to comment on my (perhaps insightless) thoughts? What are the failure statistics on HD's, do you work for a hard disk company, if so do you have access to that kind of information?

 

[ReiAyanami]

seagate says they plan to have 1 terabyte/sq inch HDs in 2010. to put that in perspectice you can fit the library of congress on a quarter. or have a 1TB ipod or muvo

 

[MadAd]

Thats the same kinda claims that were made for fluorescent multilayer disk (FMD). It was *supposed* to use incoherent light which could stack data in a 3d fashion because of the way the light made a kind of smudge (instead of being a 2d reflective surface like today). They were predicting that with a block of said material that you didnt need spinning disks anymore and the data capacities were up in the Tbs somewhere.

what happened to that? it seems to have disappeared off the face of the earth for whatever reason, and it sounded so feasible too.

 

[MadAd]

Originally posted by: Mark R
The 90 GB version is just over $27k

arrghhh *scream* how much? I guess santa wont be bringing me one of those anytime soon.

 

[Mik3y]

actually, my uncle is a very well-known comp engineer who works in san jose. hes been telling me some of the prototypes he's seen that companies are working on. they include the glass cube. by fact, its very realistic. if you have seen the glass cubes with the holographic pictures laser etched in them, that is cheap and primitive example of the glass harddrive. harddrives can only get so fast wiht rpm.