Seagate Reaches Terabit Areal Density Milestone

[back_by_demand]

[citation][nom]bin1127[/nom]Porn; movies; all the same. No plot and bad acting.[/citation]
So Police Academy is a porn movie?

 

[cknobman]

I cant wait for hard drives so large with read speeds so slow it takes over a week just to get the data off the drive.

YAY
/sarcasm

 

[CaedenV]

[citation][nom]stephenkendrick[/nom]I know it's of little concern in a desktop/server device, but how much power is required for the heat assist element of this mechanism? In turn, how much heat is developed and must be dissipated? Presumably it's very small, but if not, the extra battery drain would not be appealing in any mobile device. Couldn't find any data on-line - anyone have any ideas?[/citation]
This is simply not for portable devices.
The movement is that your computer/laptop/phone will move to flash based media (if it is not already there), and then you will have cloud or network storage that will use these drives to store your bulk media. With wiGig and 10Gbps ethernet right around the corner for home use I think we will see more and more people moving away from large internal HDDs, and more to fast SSDs for their systems, and then having these types of drives in a media server for the house to store all of your music, movies, and documents.

Personally I am getting ready to build a 4-6TB server (4x2TB drives in either RAID 5 or 10), and I am wondering how I will fill it up. I do video editing on the side, and even with my HD camera I would be hard pressed to fill that thing up by the time the drives are old enough to replace (5 years), and most projects I do are still on SD. I am not saying I will not find a way to fill it up, because I am sure my usage patterns will change to fill the space given, but as of right now I cannot think of how I will fill 6TB, much less the 60TB mentioned in the article. Maybe I will move all my software CDs to ISOs finally... but even then I don't think I will fill up everything.
The funny thing to me is that my server build is put on hold at the moment because I feel that the cost of decent 2TB drives are too high right now, and I don't want to spend $600 for 8TB (4-6TB usable after RAID), but I am chomping at the bit to find a good 240GB SSD for $240... but then again I am happy with the space of my HDDs at the moment, but not happy with the speed of them, and the server is more of a convenience, while the SSD is becoming more and more of an issue of necessity.

 

[TeraMedia]

I wonder about the reliability of the first generation of these. Will repeatedly heating and cooling spots on the drive (e.g. for Virtual Memory with pagefile.sys) cause thermal stress that will induce surface cracks? Also, what kind of write speed do these achieve? I mean, how quickly can they heat up a spot on the drive enough to alter the magnetic field susceptibility of that spot?

That said, I yearn for a return to cheap storage, and this may evoke that return in the market. Before the Thai flooding, we were approaching 3 TB / $100, or 3.3 cents per GB. Even today you're lucky if you can get 2 TB for that amount, (5 cents per GB), and the talking heads are saying that the shortage will continue through this year.

There is one consideration to factor in regarding GHz20's statement. Yes, you can put these massive disks in RAID arrays, but the rebuild time goes beyond mere annoyance, beyond even impracticality. There was an article on Toms a couple of years ago that discussed the significance of bit error rates when rebuilding massive arrays. And the problem was that there was a real possibility of a data error occurring during a rebuild, which would pose problems for array recovery. While these drives might not yet be at that point (assuming they have a similar BER to PMR), they are knocking on the door.

In my opinion, it's time for the HDD industry to migrate the desktop case designs from 3.5" drives to 2.5". The server world has already begun to do this. SSDs are almost all 2.5". Is it better to have 4 x 2.5" HDDs at 2 TB each in a RAID 5 (6 TB) or RAID 10 (4 TB) array, or 2 x 3.5" drives at 6 TB each in RAID 1 (6 TB)? Which will yield better throughput? Faster rebuilds? Better airflow in the case?

If you're storing 6 TB of data, you need a better RAID recovery strategy than simply, "plug in drive, wait 3 days." That's not a strategy. That's a failure waiting to happen.

 

[jaber2]

So, is 60TB is the limit? we can't pass that? are you sure of that or you just waiting for IBM to announce 10000TB limit?

 

[eddieroolz]

While I'm not a fan of Seagate, this is an excellent breakthrough.

 

[zaznet]

[citation][nom]jay2tall[/nom]Who says SSD's are the wave of the future? haha. I will never have to delete anything ever again with 60TB.[/citation]

640K ought to be enough for anybody. (famously miss-attributed quote)

This sort of statement has always proven to be very short sighted.

 

[rosen380]

"So, is 60TB is the limit? we can't pass that? are you sure of that or you just waiting for IBM to announce 10000TB limit?"

...
"With a theoretical areal density limit ranging from 5 to 10 terabits per square inch, capacities will likely reach to 30 TB to 60 TB for 3.5-inch drives and 10 TB to 20 TB for 2.5-inch drives."

They see a cap in how densely the data can be stored using this technique that corresponds to about 60TB on a 3.5" drive.

 

[ramon zarat]

[citation][nom]loksfox[/nom]sounds interesting, how about also increasing read/write speed ? it would take ages to backup all that data[/citation]

Transfer speed is directly proportional with areal density. The more densely packed the data, the more data fly under the head for the same constant amount of time. IE: If for a 7200RMP drive, the head travel 1 inch in 1 second (hypothetical time, just for the example) at a density of 250Gb per inch, you'll effectively transfer 250Gb per second. Double the density to 500Gb, and the head will go across 500Gb of data over the same second and the same 1 inch, effectively doubling your transfer speed. A 60GB drive with such density would be amazingly fast, even at the same 7200RPM.

 

[rosen380]

Well, that assumes that these drive will work off of a tech that can run of the RPM of current HDDs.

If these need to run at, lets say, 720 RPM, then the density would have to be 10x of a standard drive for the speed to be the same, and the projected initial density would only be around 2.5x

 

[gm0n3y]

For those talking about 3D porn, its already here:
http://www.youtube.com/watch?v=eYv3Yj6CMro

I haven't seen the actual movie, but the trailer says its 3D.

 

[__-_-_-__]

I'm more worried about speed...

 

[pedro_mann]

[citation][nom]cknobman[/nom]I cant wait for hard drives so large with read speeds so slow it takes over a week just to get the data off the drive.YAY/sarcasm[/citation]
You realise that higher density gives you a bigger speed boost than rpms right? I would like one of these drives just for the read and write speeds even if it were only 10% full.

 

[back_by_demand]

Thing is as HDDs have got bigger so have the media we use, at one time all I had was a 500GB drive and quite a few TV shows, you know the ones, 350mb each standard AVI, could get a whole season of 24 for about 7Gb and movies were those 700mb ones, designed to fit on a CD (so nasty)
...
Now I have almost 10Tb of storage but I have found that instead of having vast amounts of stuff I still have a relatively modest collection as I now get HD MKV movies or TV shows, easily 4 times bigger in size so all the storage is doing is playing catch up to the files sizes that are getting bigger too
...
By the time these drives get to the 30Tb range we will probably have 4K2K video as standard for TV and Movies and the the whole grisly circus will keep on going up and up.

 

[dalauder]

I was concerned about read errors corrupting drives or, especially, RAID arrays. But then I looked up the 1 error per 12.5TB thing people talk about and that's for 10^-14 errors per read operation. Most newer drives are 10^-15 errors or 1 error per 125TB of reads.

So in the case of my three drive RAID 5 Samsung Spinpoint F3 1TB array, that means there's a 2TB/125TB = 1 in 62 chance of a read error when rebuilding my array. That's much better than the 1 in 6 I'd heard before--an acceptable risk IF my most important photos, etc., are backed up elsewhere.

Fault tolerance seems to improve with time, so if a standard SATA drive improves to 10^16 fault tolerance, then a 60TB drive isn't really THAT dangerous. Just like a 1 in 20 (60TB/1250TB) chance of losing your data on a RAID 5 rebuild and MUCH smaller chance in a RAID 6 rebuild (assuming only 1 drive goes down).

Either way, I'm going to backup my RAID 5 to a single drive and external.

 

[dalauder]

[citation][nom]back_by_demand[/nom]By the time these drives get to the 30Tb range we will probably have 4K2K video as standard for TV and Movies and the the whole grisly circus will keep on going up and up.[/citation]The thing is, with HD content, the vast majority of people can't tell a difference between 720p and 1080p unless it's on a computer monitor. I still feel like it'll be a while before any content over 1080p @ 60fps (new movie framerate, THANK YOU Peter Jackson) is prevalent.

I mean, I have 20/10 vision, which is actually pretty exceptional, and I'm happy watching 720p from my couch. Now I'd be happier with 1080p, but I don't think more than 1 in 20 people could tell a difference between 1080p and 4K2K video from 7+ feet. Now if I have a 100" screen, my opinions on this may change. But first I'll need a bigger living room.