Feeds

back to article 6TB - big? Pah! Seagate plans to put out 8TB and 10TB MONSTERS

We got all excited about Seagate's latest 6TB drive but, really, we haven't seen anything yet. It has been revealed that the spinning rust company is anticipating 8TB and 10TB drives later this year. This info came out in the company's third quarter results earnings call, where chairman, president and CEO Steve Luczo was talking …

COMMENTS

This topic is closed for new posts.

I just hope the MTBF exceeds the time required to format one of these.

28
1
Bronze badge

Darkness

"spinning dusk company"? Tell us more about the underworld now producing a product to compete with "cloud" storage.

3
1
Coat

Re: Darkness

You really do not want to access the technical documents about such an underworld - messing about with the (blue) Prints of Darkness is dangerous business.

Mine's the one with the Necronomicon in the pocket...

5
0

One wonders at what point does the storage density start to approach that of a black hole, but given the MTBF of some drives (actual, rather than calculated or indeed claimed) the idea of all your data disappearing forever into an effed up HDD is already extant.

4
1
Anonymous Coward

I'd love to see how replacing some 3TB drives with these 10TB monsters in your typical RAID 6 deployment will alter mean time to failure estimates on some of the big players kit out there.

3
0

I think that RAID arrays are going to become a thing of the past; as drive units get bigger and bigger (and perhaps less reliable, perhaps not) you are going to see people migrate off RAID and onto the fancy data redundancy across disparate drives systems. These will cope better with drive failure and huge drive rebuild times because they are designed to do that. At 10TByte in classic RAID you're looking at 1-2 day rebuild times, ignoring whatever bandwidth is lost to the application.

Also I can see a lot more journalling going on rather than overwriting the same old block repeatedly; sort of treating the HDD like a NAND.

Maybe not though ... I'm no storage guru!

3
2
Bronze badge

Mean time to failure

It's not the mean time to failure that's the problem - it's rebuild time after a failure.

1
0
Silver badge
Linux

Re: Mean time to failure

> It's not the mean time to failure that's the problem - it's rebuild time after a failure.

...which turns out to not be a problem really. If you're really paranoid, then have a 2nd parity drive.

4
2
Silver badge

If you crunch the numbers, you start seering that raid6 is enough for the moment.

However I'm paranoid and use raidz3 already.

2
1
Anonymous Coward

Re: Mean time to failure

"If you're really paranoid, then have a 2nd parity drive." No. If you are not at all worried about your data then do that. If you are worried, RAID1. If you are really worried, RAID 11 FTW! If you are paranoid, RAID 111. If you are really paranoid RAID 1111.

0
0

RAID6 even with 4TB drives gives rebuild times that are bigger than one would wish. Fortunately there are solutions to this, Google for "dynamic disk pools" to see one fix, it is like 8D+2P RAID6 but is done at the block level. Of course this is for large numbers of drives.

0
0
Anonymous Coward

Not sure why you've been downvoted...

.. because the overlapping tracks also coming this year/very soon mean that writes will be preceded by a read and will happen in blocks of tracks.

This is exactly how SSDs work, so your concept of journalling/caching etc is probably not far from the truth.

0
0

"At 10TByte in classic RAID you're looking at 1-2 day rebuild times"

Umm, no. 2/3/4 TB are already demonstrating this. So 10TB will be (at the same speed) linearly worse, i.e. 4/5 days plus

0
0
Anonymous Coward

Don't care where you come from...

...that's a lot of... backup.

I'm thinking that's enough to see all of my favorite movie franchises, in a single disk. Think Star Trek, Star Wars, The Hobbit, LotR, in full HD glory, without any need to get up and change disks. Add in a lot of matching series...

Or even better, a $Xbox or $PlayStation upgrade, and never have to uninstall a game. That's more than I need today.

But yeah, a RAID array with these suckers will take some time to build / recover. Yikes.

1
0

Re: Don't care where you come from...

Yeah, it is - but where I come from, having the ability to store _hundreds_ of TB of data on a daily basis is normal business.

It's a simple attribute of modern CCTV systems, and a NAS full of 4TB drives for storing non-essential footage 'for a bit longer' would be upgraded to 10TB drives in an instant.

They're a greedy lot, those security types!

0
0
Silver badge

Net cost of saving data...

...seems to approach zero as the density goes up up up. Maybe that is why there are services that offer "free" storage for such things as email.

The problem is (and will always be) how to index the data. Otherwise it is just a pile of #$%@.

The retail price of 1TB of data is <$100, and it wasn't too long ago when 1TB was a pipe dream that occupied acres of space. I can go back 45 years, and I don't want to think about the changes (they have been pretty dramatic!).

So, what is a TB between friends!

0
0
Anonymous Coward

124TB drives?

I'm hearing amazing things about Radial Arc Magnetic Recording and Longitudinal Arc Magnetic Recording coupled with Density Normalization / Density Grouping.

Move over HAMR - I want to store my music collection on a RAMR-LAMR DN/DG drive

11
0

Re: 124TB drives?

> "- I want to store my music collection on a RAMR-LAMR DN/DG drive"

You are Barry Mann and I claim my five pounds.

2
0
Silver badge

"And so you are not going to kind of be able to call up and say, by the way I need an extra 500,000 eight TBs I forgot to order"

Even the concept of someone being able to casually forget about ordering Four Exabytes of storage because it's such a trivial amount is somewhat mind-blowing!

2
0

Increasing the wrong thing

4tb is enough in 1 drive, too much when the hdd dies. Maybe it's time to put a freeze on capacity and work on the reliability.

7
1
Anonymous Coward

Re: Increasing the wrong thing

I want to upvote your comment all the way to the manufacturers.

0
0
Anonymous Coward

Re: Increasing the wrong thing

Capacity is cheap though, it's only extra spinning rust (add platters), reliability costs because it requires better engineering, closer tolerances, better quality rust etc.

You can have reliability but I bet you won't want to pay for it.

2
0

Re: Increasing the wrong thing

I am willing to pay for it after having cheap WD green drives die on me, now im onto the WD blacks.

1
0

What about 2.5" drives

I'd love to see 3TB+ but right now it's tricky to even get hold of a slow 2TB 2.5" drive...

0
0
Bronze badge

Re: What about 2.5" drives

Also when did 3.5" become enterprise standard? From my experience it seemed that those enterprises serious about having lots of reliable disk went for 5.25" drives.

0
3

Re: What about 2.5" drives

3.5" became enterprise standard when manufacturers stopped building 5.25" drives in quantity. ;) Quantum did briefly (mid to late 90s) bring out the "Bigfoot" range of drives going back to 5.25" standard (one of the earliest drives to exceed 1 GB).

2
0

Re: What about 2.5" drives

Given that the last hard drive produced for the 5.25" form factor rolled off the line 16 years ago, was by then already just a curiosity, and had the astounding capacity of 47 GB, I'm genuinely curious about when and where your experience was. Is there really someone still running these relics? And they're still working?

3.5" became enterprise standard about 20-25 years ago, fwiw.

3
0
This topic is closed for new posts.