Seagate expects a significant increase in disk capacities this year and reckons it's got the jump on the competition. Stifel Nicolaus' Aaron Rakers was present at recent Seagate analyst briefing and has posted a note to clients explaining Seagate's view of the world. First of all Seagate is unfazed by the supposed coming tablet …
Make that 4GB of NAND cache on the Momentus XT.
For consumers, bigger drives are mostly pointless
Except for a few people who want to archive video, it seems that multi-terabyte harddrives are pretty much pointless in the consumer space. Speed and smallness of the drive seem bigger selling points. So while SSDs may be suply-constrained for now, it must only be a matter of time before they eat Seagate's lunch. Also, a line of 1.8" drives might be a good idea.
> Except for a few people who want to archive video
Which is pretty much anyone that bothers with home video.
Video files have always been big. Better compression in more recent devices have shrunk down file sizes a bit but the move to HD has undone that a bit. HD video of any sort quickly consumes all available space.
Instead of all of that being on DV tapes, it will be on computer disks of various kinds.
"Which is pretty much anyone that bothers with home video."
Define 'bothers with'. If I want to record something from the TV I do it on the HD+ box. If I don't want to do it there I watch it on iPlayer/4oD/&c.. If I want to watch a DVD I either rent it via HD+ box or buy it (cheap) online.
I don't currently have a use for anything larger than ½TB.
I suppose if you record amateur porn at home maybe this kinda drive will be a lifesaver, but who else is going to use it?
"but who else is going to use it"
Uh, gamers ?
Let me check the folder size for some of my favorite games :
Battlefield 2 takes 9.9GB
Crysis Warhead takes 3.7GB
Freelancer takes 4.1GB
Hellgate London takes 6.8GB
Lord of the Rings Online takes 12.9GB
Starcraft II takes 8.7GB
Company of Heroes takes 3.9GB
and my Steam folder, with dozens of games, takes up 95.5GB.
Sure, this only takes 145GB in all. You might think that a 250GB disk should be amply enough. Yes it would - for now. But neither you nor I know what games I will add to my collection in the future, and I'll be darned if I want to start worrying about disc space again like I had to ten years ago (when game size on disk was counted in MB, not GB).
I bought a 1.5TB disk for the express purpose of installing my games on it. Yes, I have more than 1.2TB available on it, but that means I have space to install a lot more games. More importantly, it means that defragging is not an issue, and the disk can manage its sectors without problem either. Plus, I have peace of mind and room to grow.
Finally, the disk only cost me 89€. Hey, that's the price I paid for 320GB back in 2007 !
Cretins R Us
They just can't get by the "bigger is better" obsession. Bunch of little boys standing in a circle trying to see who has the biggest hammer. Too bad they can't seem to understand they need to be good before they start worrying about size.
IBM and Hewlett-Packard
..proved that rotating disks can be very small (HP) and resist high-g shock (IBM). Now all seagate has to do it to develop a matchbox-sized HDD and have an accelerometer wired to the drive to get the read/write heads off the platters when there is an exception state of acceleration.
More radically, develop a drive with heads moving on two support rails up and down the center of the rotating platter. These drives should be able to withstand much higer g rates.
Size doesn't matter
"First of all Seagate still expects its hybrid Momentus XT, the disk drive with a 4MB NAND cache, "
Don't want one of them, got any with 4gb cache?
"while the 3TB, 5-platter Barracuda XT goes up to 3.9TB - surely that would be pushed to 4TB though."
they will surely market it at 4 TB, whether it is or not.
Does Anyone Remember?
My first PC, an Amstrad 1512, came with a whopping 10Mb hard disk. And you could install lots and lots of different software programs on that space. Whereas today 10Mb won't even hold your plumpers URL list.....
You kids today
Your first PC had a hard drive... (no where's the avatar for Thomas Watson Jr?)
35% failure rate for enterprise SSD? That seems more than a little OTT
Also, on the top of page 2 the Momentus XT has 4 GB of flash not 4 MB.
No that seems low
Strictly speaking, you expect all hardware with known "wearing" factors to fail eventually - so presumably the failure rate would be closer to 100%.
The 35% failure rate is meaningless without the timescales.
Price is king
Until the two technologies are close to one another on price point I fail to see how you can compare the two. There are two different areas of use at present. Also there is a huge market for NAND based products in small devices and corporate servers where speed counts rather than capacity. The large drive capacity sector is a different marketplace aimed at large arrays and consumerist end users.
My main concern is where size and cost is pushing reliability to the bottom of the list of priorities. I'd much rather have a reliable 2TB drive than a 3-4TB brick thanks.
Also Seagate does not have a good history recently. I've had better luck with Samsung even with having to update my firmware in their latest F4 drives. Hitachi are pretty bad too since my last order of 2TB drives were all bad.
I would love to see the back of spinning media but alas until the cost of NAND is on a level playing ground in terms of size and cost (reliability should be better hopefully I would assume). Then I'm afraid we are just stuck with it.
Capacity is not the issue...
...for most consumers, that bottleneck was breached long ago. Speed, reliability, energy efficiency, and price are the hot spots, especially outside the desktop/server market.
I wonder if the good people at Seagate could devise a device - a composite type of device - that marries SSD technology with hard disk technology AND cater for failsafe to?
That's exactly what the hybrid disks are all about - a moderate SSD-like cache in front of a large rust-spinner. It does cater for failsafe - it's a non-volatile cache, so even if the power is cut, the SSD cache survives.
Of course we seem to be charitably ignoring their first attempts a number of years back that seemed to offer no speed boost which then crashed and burned in the history of failed endeavours - presumably the caching algorithms are better these days, but let's not pretend this is a "new" thing they're trying.
4GB is not enough
The problem with the hybrid drives is they're trying too little. 4GB is a great way to boost, say, OS boot times. However, a 16-32GB cache would be much more useful. And include a tool that you could, in OS, "add file to drive cache." Would be a no-brainer to let the OS get auto-cached and simply add/remove files of your Game-Of-The-Day (think the big MPQ files from Blizzard perhaps?).
The end is nigh for rotating storage
I've just bought a four 2TB hard disks for a NAS and I expect them to to be the last magnetic based storage I buy. By the time I come to upgrade them in 2-3 years, I'd expect solid state to have progressed to match that capacity, if not exceed it.
I already run buy Intel X-25M 80GB SSDs for any operating systems I install.
In 5 years time I'd expect the future to be pretty bleak for outfits like Seagate - the ability to keep ahead of SSD on capacity alone will slowly become harder and harder to maintain.
Neil, I am assuming then that 3 years ago your NAS also had 8TB of capacity, and that 3 years before that, it also had 8TB of capacity?
If not, what makes you think that your NAS will only need 8TB of capacity in 3 years time? If you need 8TB now, the chances are you shoot thousands of RAW photos, digital videos and/or have a movie collection on it. If you have any of these, then you will probably need more capacity in 3 years, no?
I could similarly say that my first notebook only had a 20MB drive; so surely all notebook drives could be flash now?
Yep, undoubtedly my storage needs will continue to increase.
But unlike 3.5" HDDs, SSDs are smaller (1.8", 2.,5") and generate much less heat (and are, arguably, more reliable since no mechanical parts to fail), so in future I could build a NAS with more SSDs to make up for any lag in terms of individual disk capacity, and the physical space taken up would be about the same as my current HDD based NAS, possibly smaller.
Today I could, quite easily, build a 6 or 8 disk NAS entirely with SSDs that would rival (in terms of capacity) a 3 or 4 disk HDD based NAS, but the cost would be astronomic (although the performance would something special!). I'd not gain anything in terms of physical space taken though, as today's NAS enclosures are (understandably) designed for use with spinning 3.5" drives, and with all the thermal management requirements those hot drives require.
Once SSDs become "large enough" and - perhaps more importantly - cheap enough, magnetic storage is dead. And NAS enclosures will become a lot space efficient. :)
I like Seagate. I really do. Until very recently, I wouldn't even contemplate using any non-Seagate drives, in fact.
But a quick check at New Egg, for example, shows only 2 internal 3TB drives: WDC Green and Hitachi 7K3000. If memory serves, the situation was similar with the 2TB drives, with Seagate trailing by several months.
So, where's Seagate's alleged technology lead?
When I was 18
My first job was at IBM where we built Kestrel drives; 38kg in weight, 375MB capacity. By the time I was 21 they'd grown to 3.5GB and shrunk to the size of a cigarette packet.
I just thought I'd share that with you.
Yeah I saw some motion pictures, and just after Jesus and the dinosaurs were the first - the VERY first hard drives ever made... in Australia of course. (Americans and Brits = retards).
They were like 3 feet across, brown and about 2 or 3 inches thick and held 24???? ummm 1.2 Meg or something.
Brilliant - manhole cover sized hard drives...
2TB to 2.6TB?
From a user point of view though, this doesn't seem like much a difference.
For me, I like the Western Digital MyBook range, and I've got a bunch on 1tb ones from when they were about £90, and then when the 2TB ones came down to about 90quid last year, I got a couple of those.
So I'm not likely to buy any more at 2.6gb. I'll be waiting for when the 4tb external drives come down to 90quid before I next buy disks, and by that time, perhaps the SSD tech will be affordable?
So while it's great news for anyone buying a new machine this year, or just now thinking about upgrading their internal disks, I don't see that Seagate are going to get a massive increase in sales, this increase in capacity is not enough to encourage people to upgrade their storage, they'll just make do with what they have for the moment.
4MB NAND cache?!
Really that tiny? Drives have bigger RAM caches than that today. A 4GB cache might have some benefit.
"...total notebook HDD capacity of 95EB (exabytes) shipped in 2011 with 69EB in 2010."
"This left a pathetic 0.77EB (770PB) for notebooks, which would have allowed around ten per cent of notebook HDD storage to be replaced by SSDs last year."
"...boost in NAND fab capacity this year, maybe enabling 2EB to be used for notebooks
Its even better than this for the spinning metal mob as .77 is in fact slightly more than 1% of 69 not "around ten per cent".
Working with the 2011 estimates NAND production will still only be able to supply about 2% of the forecast 95EB. Still the writing is on the wall for rotating disks in the laptop/notebook/netbook market maybe 5 years before its all flash.
probably my calculating skills have left me on Monday morning, but I thought that 0.77 EB on a total of 69 EB is about one percent instead of ten...
The fanbois are humorus
It's always a good laugh to see the pissfests by fanbois. If it ain't Seagate vs. WD then it's conventional HDs vs. SSDs.
For most folks increased HD density/capacity is of little significance except for servers. I'm not a fan of Seagate or WD. Both have made good and lousy products over the years.
As far as SSDs go they still need a quantum leap in reliability before they will be acceptable for mainstream use. The reported 35% failure rate is probably conservative.
Quantum leap in reliability
Shirley if SSD's undergo a quantum leap in reliability, there's an equal chance they'll become even more unreliable?
Longevity of data
is really what worries me. We've had a relatively golden age for the last 15 years or so, where any media that you could write to is probably still readable now. I certainly can still read CDs that I burned in the 90's, although it probably depends on how they have been left in direct sunlight.
I recently had a requirement to read some 'spinning rust' from systems I ran in the same time frame (one was a 1.1GB Quantum disk, another was a ~860MB Seagate disk), and the data was still there, still readable.
I have, however, found that older media, particularly floppies (5.25 and 3.5) from the same period or older are very much more a problem, with an almost 50% failure rate of those that I tried (including a whole set of precious and irreplaceable BBC micro disks).
I worry about how long an archived MLC or even SLC flash drive will remain be readable after being put on the shelf. I have already had various flash cards fail on me, which is not a problem for cards used for transient information (holding photos until they are loaded onto a computer, or copies of ripped music or video that is also held elsewhere). Throw them away, and buy another one, reload the data if applicable.
But this would be more of a problem if flash was being used as the primary repository of the information.
I'm not sure that disk is the correct solution either (especially as old style SCSI is pretty much dead, and EIDE interfaces disappear from new systems, replaced by SATA, and SAS and other serial technologies), but I predict that it will be more useful for archive storage than Flash memory. I'm purposefully ignoring tape, as this is now far too expensive for ordinary people, even though the remaining tape and drive manufacturers have a roadmap for data longevity.
Looks like we are fated to continue to re-write our important data forever as we move away from media with any significant lifetime. I think we will look back to days when books were on paper, photographs were on film, and music was on vinyl, all with a lifetime measured in decades, with some nostalgia.
"Rewrite repeatedly" was always the way. Photos and text faded or got damaged. The loss of Library of Alexandria was a big deal bcos ppl forgot about that. The only difference today is the rewrite cycle period. Mind you, if you reckon on the quantity of data being produced, that's probably fair game.
More likely is that the concept of having one discrete location for data will take a tumble. As NASes become more widespread, auto-backup solutions seem likely to become more widespread too.
And the idea of of an "archived flash disk" is a non-sequitur. Flash by its design *WILL* fail over time, because charge leaks. It simply isn't an archival storage format, any more than writing in pencil is suitable for archiving. The physical medium (paper, Flash chip) may still be OK, but the data on it (writing, files) will be unreadable.
Ceramic graphite composite writing heads (pencils) on MMWD's (Manual Media Writing Drives = paper notebooks) last forever.
It's the paper that needs to be special and in special storage.
Still you could upgrade to CCWT's (Cuniform Clay Writing Tablets) and WFPK's (Wood Fired Pottery Kilns).
These have an indefinite lifespan of 200,000 years plus.
Unseriously tho we have 2 problems.
1. There is the issue of translocation of the media and media formats that retain the data - THROUGH time.... AND (think Edison wax cylinder then shellac records etc..)
2. Who in their right mind wants to keep all yoru precious shit after your dead and buried?
Your own children? Probably.
Your grandchildren? Probably not.
Your Great Grandshildren? "Who the fuck was he?
Grand Museums of the world - yeah your hard drives and those of 200 billion other people.
Sure... even tho many fillums of the celluloid and nitrate variety were once considered great works of art.... most of them have either gone to the tip, rotted away in their cases or caught fire.
Few if any of the overall amount ever made, survived to make a restoration job from, same goes for all the 78's ever made of Bobby Bell the Tenor of Bellarus in 1925 and the Clarinette Sonata in C#, and everyone else who ever scraped the wax.
Still maybe the magic sticks on the wall can make you or your memory live on forever.
Just ask the guy who rides the dinosaurs.
Actually - who can I bribe to get into this eternal life thing?