Consortium Roadmap Shows 100TB Hard Drives Possible By 2025

Lucas123 writes An industry consortium made up by leading hard disk drive manufacturers shows they expect the areal density of platters to reach 10 terabits per square inch by 2025, which is more than 10 times what it is today. At that density, hard disk drives could conceivably hold up to 100TB of data. Key to achieving greater bit density is Heat-Assisted Magnetic Recording (HAMR) and Bit Patterned Media Recording (BPMR). While both HAMR and BPMR will increase density, the combination of both technologies in 2021 will drive it to the 10Tbpsi level, according to the Advanced Storage Technology Consortium (ASTC).

Comments

Luddites everywhere

[Anonymous Coward]

Clearly now that we have 3D printed a piece of cake frosting in free fall, we will 3D print any kind of technology at home in our living rooms.

I'm gonna need more porn torrents

[Anonymous Coward]

n/t

But what about

[Anonymous Coward]

http://hardware.slashdot.org/story/14/11/25/2027220/how-intel-and-micron-may-finally-kill-the-hard-disk-drive

Re:But what about

[houstonbofh]

Doesn't matter. It is ten years out. This prediction won't even be on Archive.org anymore by that time...

Re:But what about

[MrL0G1C]

Flash memory gets less reliable as the density increases so it doesn't look like we'll be seeing 100TB flash drives with SLC/MLC technology.

Re:But what about

[nospam007]

"Doesn't matter. It is ten years out. This prediction won't even be on Archive.org anymore by that time..."

Did the predicter claim his right of his bad predictions to be forgotten?

Re:But what about

[Russ1642]

You think your assumptions about flash memory will be true in ten years? LOL

Re:But what about

[Anonymous Coward]

Process shrinks are already becoming INSANELY more difficult- dummy- and Flash already uses the highly vertical fabrication methods FinFET is only now introducing for general computer chips.

As Flash cells shrink, their reliability falls off a cliff. On-the-fly error detection/correction requires every more chip space- and for MUCH larger conventional logic transistors. Current flash works by voiding entire blocks of memory when errors occur- creating really really scary failure modes compared to the HDD.

Flash is best seen as a convenient, disposable form of memory- used as cache between RAM and HDD- or as a convenient, portable method of moving data in physical form.

Flash replacing general HDDs is industry hype- sold to credulous investors with the assistance of loud mouthed know-nothings on tech forums like Slashdot. Flash as a HDD replacement has revolutionised certain mobile devices, although the bigger of these would still be better off with a flash cache before an internal small factor HDD. The really light, thin devices can only use Flash, however.

The synthetic manipulation of end-user RAM prices (which are easily FOUR times higher today than what they should be) makes Flash seem a more exciting tech than it actually is, because currently Flash sees the regular price erosion that once was commonplace with RAM. What is true is that is you use a REALLY lazy and badly designed file system, whacking in a large (and expensive) SSD drive to replace your HDD can give massive benefits. But with a better file system, a MUCH smaller amount of flash acting as level-5 cache (RAM is level-4 cache) would give almost the same performance boost.

Re:But what about

[Anonymous Coward]

You think atoms will shrink in ten years? LOL
http://www.youtube.com/watch?f...

Re:But what about

[MrL0G1C]

+1, great video.

Re:But what about

[GigaplexNZ]

They've stopped shrinking the process (they're actually going bigger again) and are just layering NAND vertically for extra density. We've still got some way to go with reliable high density flash memory.

Re:But what about

[toddestan]

Why not? You simply stick more flash chips in the drive. It's not like it's a mechanical device with moving parts where it gets impractical to keep stacking platters.

Re:But what about

[Anonymous Coward]

Wishful thinking by Intel and Micron if they are really saying such nonsense.

I work in technology development upstream for both Semiconductor and HDD industry. SSD is the technology that has future limits. HDD has plenty of future scaling. That's not the dogma but it's the truth based on physics, engineering and empirical realities. That's not what pundits generally pontificate from however.

Re: But what about

[Anonymous Coward]

Nonsense, drives at nanometer track widths have to figure out how to build motors that can drive them without missing tracks due to vibration. The real hard part ahead isn't the shrinking of aerial densities it's the mechanical challenges with vibration that's going to make the 100tb drive a hard sell. And the magnetic guys gets no benefit from consumer since the have abandoned magnetic drive technologies (except the enthusiast class of consumer). The big success for flash densities comes from the mobile segment of consumer and there is quite a bit of money going into that research. And Lea than five percent of the flash output goes into enterprise (is largish) ssd's. The rest hits the embedded market like cell phones and tablets.

Now the issue of rber and size is a problem for the mag drives just as much as the flash guys. Once your past a tb you pretty much have to go into dual parity for protection. At fifteen tb you are into triple parity and from there it gets worse. Flash can alter the layouts and since reads are free they can do far more recursive protection schemes than a drive can do without retorting to write buffers (but then you have to bet on large block writes which will mess up locality of reference which leads to metadata caching, which then leads to hybrid key/value drives with block access and boy is it messy enough yet? And while all this is going on only the buffet of enterprises are the customer base so they get to pay higher and higher r&d cost that came be amortized over 6 billion consumers.

Mechanical computing and storage is dead. Silicon is winning. Disk is that last bastion of mechanical computing followed by tape and it's dying on the vine. Welcome to the future...

How about transfer rate and reliability?

[mlts]

MTBF and transfer rate numbers are boring... but those can be just as important, if not more, than the drive's capacity.

With high capacity tier 3 drives, one reason that RAID 6 (or a RAID 50 setup with tiers/groups of disks) is used is because it can take days to rebuild a blown drive. If drives continue to have larger capacities, but I/O stays the same, then we will need to add more parity drives to RAID arrays to support multiple drive failures and still keep the data accessible, better algorithms that run in the background to detect (and fix) bit rot, and bigger/smarter caches.

Maybe this is just me, but I'd rather see drives with double the MTBF than double the capacity. I can always add more drives and arrays. A failed disk will cost time no matter what, even if it is just walking to the server room, pulling it out and replacing it with a spare. For non-enterprise customers, a failed drive can be catastrophic since not many users have RAID arrays for protection.

Re:How about transfer rate and reliability?

[Anonymous Coward]

The MTBF of a typical non differentiated SSD modestly exceeds that of the typical HDDs found in consumer level products available to the average consumer.

The seek time of SSDs is significantly higher than HDDs however, as SSDs have more complex sector location algorithms.

Re:How about transfer rate and reliability?

[rrohbeck]

WTF? HDDs have seek times in the milliseconds while total access time for SSDs is in the microseconds.

Re:How about transfer rate and reliability?

[rrohbeck]

There was an article recently (can't remember where) that made the case that with slowing density increase the lifetime of HDDs has to increase because you're not going to replace them after two or three years anyway because the next generation is so much better. Much better MTBF is clearly possible - just look at HGST versus the rest in the Backblaze reports.

Yup, in large arrays the trend is to go beyond RAID6 - see e.g. NetApp's DDP. Too bad there's so little technical info available about it.

Re:How about transfer rate and reliability?

[Billly Gates]

Not to mention the grandparent has a backward idea of reliability too. Though the very newest SSDs are significantly more reliable you can ask any shop owner what he thinks of SSDs. They do die more frequently than a disk based on customers bringing them in. Linus Torvalds lost his too out of nowhere with no warning.

I have a raid with some SSD's on the system I am typing this on but I have my swap and personal data/profile on a mechanical disk still for this reason

Re:How about transfer rate and reliability?

[Anonymous Coward]

Not quite. SSDs have electromechanical electrostatic bridges that require positive charge dissipation. These microscopic bridges each require an average of 100 microseconds. With a 100TB SSD in 10k clusters we would have 64 address bridge transitions required per sector access. With a resulting average of 32 bridges, we have 3.2 ms per access. Not large but not insignificant.

Re:How about transfer rate and reliability?

[Zontar The Mindless]

Torrents, my good man. Torrents.

Re:How about transfer rate and reliability?

[rrohbeck]

I have several VMs on my system that each need at least a half TB to be useful (better a full TB.) A local cache of our SVN repo is again a couple 100 GB. Next, backups of my other systems, each a couple 100 GB at least. Next, a mirror of my server at work - nearly another TB or so. And a couple movie downloads that I haven't watched yet. Voila, over 5TB.

Re:How about transfer rate and reliability?

[Anonymous Coward]

NO. 100TB / 10KB clusters = 10 billion clusters on the SSD. Since most SSDs use SAMR3 addressing we only require 34 bits in the cluster address, At 17 bits average address resolution we have an average of 1.7ms for address bridge resolution. Once the cluster location has been ascertained we may naturally access the cluster in 10us.

It is therefore determined that an average access will require 1.701ms.

You are welcome for the correction.

Re:How about transfer rate and reliability?

[rrohbeck]

You forgot the time to charge the flux capacitors.

Re:How about transfer rate and reliability?

[Anonymous Coward]

HDDs and SSDs do not use flux capacitors. Flux capacitors are fictional components from the movie Back to the Future.The movie was before my time but Wikipedia was friendly enough to tell me.

Re:How about transfer rate and reliability?

[Anonymous Coward]

I just downloaded a siterip of everythingbutt at 357GB. Also eyeing a 350GB Japanese Bukkake torrent. Those are just niche btw, theres 1TB+ mainstream porn torrents out there. 4k/8k will really be the death of us.

Anyway, if lesbian rimming isn't a good enough reason for you, Idk what is.

Re:How about transfer rate and reliability?

[Anonymous Coward]

It's called a joke. Look that up too.

Re:How about transfer rate and reliability?

[Anonymous Coward]

Flux capacitors do exist:

http://www.tvtechnology.com/default.aspx?tabid=204&entryid=3607

Re:How about transfer rate and reliability?

[Solandri]

Transfer rate is proportional to areal density. If the areal density increases 10x, then for each rotation of the platter 10x as many bits pass under the read/write heads, and (sequential) transfer rate increases by 10x as well.

The problem for HDDs is and always has been random seek times - the time it takes to move the read/write heads to a new location and wait for the proper part of the platter to spin underneath. Look at this 7200 RPM HDD reivew from 2003. The sequential read/write speeds (45/27 MB/sec) are about a third what a modern drive gets. But the IOPS is exactly the same because it depends platter rotation speed.

Re:How about transfer rate and reliability?

[fnj]

1) If one is not using ZFS with RAID-Z2 (double parity) or RAID-Z3 (triple parity), one is an idiot. I think we're in basic agreement, but I'm just emphasizing that no crude RAID technology can compare with ZFS.

2) You do understand that MTBF has nothing to do with design lifetime. Right?

Re:How about transfer rate and reliability?

[fnj]

As long as only a single head is active at a time, reading a single track at a time, transfer rate is proportional to LINEAR density, not AREAL density.

Re:How about transfer rate and reliability?

[Anonymous Coward]

What the hell do flux capacitors have to do with SSDs? Can flux capacitors power SSDs? On spacecraft perhaps?

Re:How about transfer rate and reliability?

[s.petry]

Some people like to save their pr0n for watching later

Re:How about transfer rate and reliability?

[Anonymous Coward]

Nothing

Someone responded to a serious post about the electromechanical bridges used in SSDs by suggesting use of a flux capacitor.

Nothing to see here.

Re:How about transfer rate and reliability?

[Anonymous Coward]

From the website cited above:

For whatever reason, people have been using the term flux capacitor for years as a synonym of some technology that’s really cool — but doesn’t exist. However, perhaps Doc Brown was just ahead of this time.

In mid-August, a paper published in Nature Nanotechnology describes a new kind of capacitor. The device is constructed of concentric shells of graphene. Graphene is typically represented by a one-atom-thick flat sheet of sp2-bonded carbon atoms packed into a honeycomb crystal lattice. Think of it as an atomic-scale version of chicken wire made of carbon atoms and their bonds.

This new capacitor is created from “onion-like carbon” (OLC) formed into a series of concentric spheres of one-atom-thick sheets of carbon. Think of the capacitor as a series of balls of increasing size, each separated from the surrounding ones by a dielectric. According to the paper’s authors, the “onion-like” shells of graphene can produce discharge rates of up to 200V per second. This is “three orders of magnitude higher than conventional capacitors,” say the authors.

Compared with thin film lithium batteries, the OLC provides less energy per volume. However, the device provides more than 10,000 times more power per volume. The authors say the OLC may find applications that require large bursts of power, long lifetimes and decent storage capacity.

These could provide great battery backups for SSD applications.

Re:How about transfer rate and reliability?

[blind biker]

Sadly, the trend with SSDs is the same: higher density without any regards to MTBF.

Re:How about transfer rate and reliability?

[Anonymous Coward]

What "electromechanical" bridges are in SSDs?? This entire thread makes no sense to me. The only inkling that the whole thing is a joke is the 1.701ms, probably referencing Star Trek.

Re:How about transfer rate and reliability?

[Anonymous Coward]

A flux capacitor could be integrated with the SSD for a self-powered storage device. If the SSD power drain is sufficiently small it could be chanced that the flux capacitor could be charged in the factory and that the SSD would be self powering for its entire life span.

Re:How about transfer rate and reliability?

[Anonymous Coward]

What does 1.701ms have to do with Star Trek?

Re:How about transfer rate and reliability?

[sycodon]

100 Terabytes.

Even I would have a hard time finding enough porn to fill that.

Re:How about transfer rate and reliability?

[Anonymous Coward]

A flux capacitor could be integrated with the SSD for a self-powered storage device. If the SSD power drain is sufficiently small it could be chanced that the flux capacitor could be charged in the factory and that the SSD would be self powering for its entire life span.

FWIW, here are the OPs calculations:

100us access timer per electromechanical bridge * (( 34 total bridges required to cluster resolution / 2 for average) = 17) = 1.7ms + 10us for cluster access = 1.701ms.

The electromechanical bridges are microengineered address resolution switches. It is quite the novel approach. Access time is fast and current draw is low.

Re:How about transfer rate and reliability?

[Anonymous Coward]

FWIW, here are the OPs calculations:

100us access timer per electromechanical bridge * (( 34 total bridges required to cluster resolution / 2 for average) = 17) = 1.7ms + 10us for cluster access = 1.701ms.

The electromechanical bridges are microengineered address resolution switches. It is quite the novel approach. Access time is fast and current draw is low.

Re:How about transfer rate and reliability?

[Anonymous Coward]

NCC-1701, duh. Really? I had to point that out? Now where are these electromechanical bridges in a SSD?

Re:How about transfer rate and reliability?

[Anonymous Coward]

What does 1.701ms have to do with Star Trek?

From Wikipedia (http://en.wikipedia.org/wiki/USS_Enterprise_%28NCC-1701%29):

The USS Enterprise (NCC-1701) is the central starship in the fictional Star Trek media franchise. The original Star Trek series features a voice-over by Enterprise captain, James T. Kirk (William Shatner), which describes the mission of Enterprise as "to explore strange new worlds; to seek out new life and new civilizations; to boldly go where no man has gone before". The ship's design "formed the basis for one of sci-fi's most iconic images".[2] A refit version of NCC-1701 appears in the first three Star Trek films. The 2009 Star Trek film, which takes place in an alternate timeline,[3] features a re-conceptualization of the original Enterprise.

Re:How about transfer rate and reliability?

[Anonymous Coward]

NCC-1701, duh. Really? I had to point that out? Now where are these electromechanical bridges in a SSD?

Read up:

http://www.csa.com/discoveryguides/mems/overview.php
http://ieeexplore.ieee.org/xpl/RecentIssue.jsp?reload=true&punumber=84
http://en.wikipedia.org/wiki/Microelectromechanical_systems

SSDs utilize a common data base to manage multiple NOR flash banks. Prevention of stray charges to nonreferenced NOR banks is mandatory, so SSDs use micro mechanical address switches to connect only appropriate NOR flash bank for an operation. Their latency is very low but not 0.

Re:How about transfer rate and reliability?

[Anonymous Coward]

And there is your mistake.

The OP is assuming full cluster resolution by the bridges. But we do not have a NOR per cluster so we only need enough bridges for NOR selection. if the 100TB has 10 NORs then we only need 4 bridges to uniquely identify the NOR.

100us access timer per electromechanical bridge * (( 4 total bridges required to cluster resolution / 2 for average) = 2) = 0.2ms + 10us for cluster access = 0.201ms.

Much better, much faster the HDD, and no Star Trek references.

Re:How about transfer rate and reliability?

[Anonymous Coward]

Thank you. I still don't think that the micro electromechanical bridges are being described correctly but the calculations look right.

Re:How about transfer rate and reliability?

[Anonymous Coward]

This compares very favorable to the hard disk drive.

Re:How about transfer rate and reliability?

[sexconker]

Fuck that shit. RAID 0 with daily backups for home use, RAID 10 with daily backups for any servers. If you're paranoid add a hot spare.
Oh, and SSDs only, even for servers. If you need more space, put on your big boy pants and fork over the cash.

Re:How about transfer rate and reliability?

[Anonymous Coward]

I am perfectly well aware of what MEMS is. I don't see any reference as to where they are used in SSDs for "prevention of stray charges". It's a baffling concept in any case.

Could you provide that link?

Re:How about transfer rate and reliability?

[Anonymous Coward]

MTBF is proportional to the bridge reliability.

Re:How about transfer rate and reliability?

[Anonymous Coward]

Read up:

http://www.google.com/patents/US6272043

Re:How about transfer rate and reliability?

[Anonymous Coward]

I am perfectly well aware of what MEMS is. I don't see any reference as to where they are used in SSDs for "prevention of stray charges". It's a baffling concept in any case.

Could you provide that link?

Read up:

http://www.google.com/patents/US6272043

Re:How about transfer rate and reliability?

[Anonymous Coward]

Sadly, the trend with SSDs is the same: higher density without any regards to MTBF.

Seems like a problem with consumers, you can still buy an SLC SSD. The problem is that since it doesn't trade off MTBF to get higher density it seems unreasonably expensive and as if no development happens on that front.
That is a problem with the way consumers looks at products and computer stores not having the option to sort by MTBF, not a problem with the SSDs.

Re:How about transfer rate and reliability?

[theshowmecanuck]

What's more, even if the seek time is very slow, once done it can go back in time so that it actually seems very, very fast.

Re:How about transfer rate and reliability?

[Anonymous Coward]

I agree for the most part. But to increase transfer rates we will need a newer, likely more expensive, platter material. Lots of people will point to ssds as the answer, but time has shown that HDDs are generally longer lasting than flash memory. Thats ignoring the cost, because it is quickly dropping to a parity point, time will tell when it comes to that.

Its seems a lot more like memristors or racetrack memory will win this race once it matures.

Re:How about transfer rate and reliability?

[Anonymous Coward]

There's nothing in there either about "prevention of stray charges", "electromechanical electrostatic bridges", or of MEMS addressing.

This isn't funny, it's just stupid.

Re:How about transfer rate and reliability?

[Anonymous Coward]

Even with RAID-Z3, you could run into trouble because rebuilding a 100TB drive is going to take a long time. Assuming the sequential read rate increases at the square root of the increase of the density, then you're talking about a 5x increase on a 3TB drive with 33x more storage. That puts it around 750MB/s peak, but you probably won't see more than 500. About a 2.5 day rebuild window. That's a long time to be down a drive.

Re:How about transfer rate and reliability?

[SuricouRaven]

I'm not sure hard drives will still be used in the same way. It's increasingly looking like SSDs will take over for a lot of things, and hard drives will take the role that once was filled by tape: Backups, archival and the lowest rank of tiered storage. In which case you won't be using RAID or even a conventional filesystem, but some form of object storage.

Re:How about transfer rate and reliability?

[drkim]

100 Terabytes.

Even I would have a hard time finding enough porn to fill that.

Ha, ha, ha, ha, ha, ha, ha. Amateur.

No, seriously, check out 'Amateur'.

Re:How about transfer rate and reliability?

[Neil Boekend]

That doesn't work linearly. It's the square root of 10: 3.16

Ten times the density means 3.16 times the amount of tracks beside each other and 3.16 times the amount of bits per track.
If the head still reads only one track that means the amount of bits it can read in a given time is 3.16 times as high.
Not 10.

Re:How about transfer rate and reliability?

[Anonymous Coward]

I just downloaded a siterip of everythingbutt at 357GB. Also eyeing a 350GB Japanese Bukkake torrent. Those are just niche btw, theres 1TB+ mainstream porn torrents out there. 4k/8k will really be the death of us.

Anyway, if lesbian rimming isn't a good enough reason for you, Idk what is.

LOL! Your wiener will be a bloody pulp of beef jerky by the time you finish "watching" all of that. Best of luck to you with that venture! And don't forget the popcorn.

Re:How about transfer rate and reliability?

[Anonymous Coward]

Regarding MTBF, did you overlook the new recording technology? They are using a HAMR. The platter will consistently break on first write.

Re:How about transfer rate and reliability?

[AlchemyX]

Thank God we have ZFS and RAIDZ3 :).

Re:How about transfer rate and reliability?

[drinkypoo]

As long as only a single head is active at a time, reading a single track at a time, transfer rate is proportional to LINEAR density, not AREAL density.

We used to have drives that read/wrote multiple heads at a time. But the drives are too dense for that. I think the last ones that did it had glass platters to reduce expansion. Now we'd need a separate arm for each platter. Some drives have two sets of arms, but that's still only two read/write tasks simultaneously.

Re:How about transfer rate and reliability?

[TheRaven64]

And how does the size of a photographs compare to a 1500 page book? Modern mobile phones come with cameras that can take pictures that are 30MB each. That 1TB is enough for 35,000 photos at that size, but it doesn't take too many doublings for it to start to feel cramped. HD video can get through it very quickly and most modern phones can record at at least 1080p now.

Re:How about transfer rate and reliability?

[Anonymous Coward]

Railgun!

Re:How about transfer rate and reliability?

[Anonymous Coward]

i used to be in storage, and when we started selling zfs-based solutions, then
we started losing customer data by the truckload. seriously, 8 years of no data loss,
then zfs and 10s of customers lost data.

the absolutist bit above does not square with the facts.

Re:How about transfer rate and reliability?

[0123456]

What the hell do flux capacitors have to do with SSDs? Can flux capacitors power SSDs? On spacecraft perhaps?

With a flux capacitor in your SSD, it will send the data before you even ask for it.

Re:How about transfer rate and reliability?

[tompaulco]

I have a 2 TB drive and I have used about 300 GB of it. So I am inclined to agree that for my purposes, 100 TB is not something I will need by 2025. I may need it by 20 or 30 years after that. After all, I never really thought I would need 300 GB. There are plenty of people who do need huge amounts of space. Some people like to play their media from their computers instead of from the DVD/Blueray. Some people like to illegally acquire movies and games. Some people like to collect porn.

Re:How about transfer rate and reliability?

[gstoddart]

1 TB has enough data for every man woman and child who ever lived to write a 1500 page book!

That's awesome and all ... but this isn't 1988, and we're not storing just text files any more. Nobody is writing 1500 page books to fill up their hard drives.

My iTunes folder is around 400GB, which includes my music and any digital copies of movies I've gotten. An HD version of a movie rings in around 5GB or so. I've still got dozens of DVDs I plan to rip so I have them on-line.

I just recently bought about 60 CDs which I need to rip, and the next BluRays I buy will include digital copies. So, realistically, I'll add another 100GB or so in the next few months.

I have about 100GB of photos, because I have owned digital cameras for the last decade.

I know people with children and HD video recorders who have massive amounts of pictures and videos. They blow through space like you wouldn't believe.

nd please do not give your niche use case in a reply

if you think my use case is niche, you're either clueless, or still only holding onto a bunch of text files.

My 2TB HD is about 50% free, and gets backed up to two separate 2TB HDs. I expect to be more like 40% free within six months.

But do I know a lot of fairly normal people without overly extreme data requirements who can blow past 1TB pretty fast.

So, I think your assertion nobody is using that much space is pretty much unfounded.

Seriously, look around you at what people are actually doing with computers. It might surprise you.

But, hey, like your namesake ... 640K ought to be enough for anybody, right?

Re:How about transfer rate and reliability?

[Anonymous Coward]

There's nothing in there either about "prevention of stray charges", "electromechanical electrostatic bridges", or of MEMS addressing.

This isn't funny, it's just stupid.

I guess you didn't read it very well!

Cheers!

Re:How about transfer rate and reliability?

[WuphonsReach]

In practice, SSDs have only 20-100x the IOPS of a similar number of spinning platter drives. Which is still a huge improvement, but not three orders of magnitude (1000x). The bigger advantage is that when you have more workers accessing the drive, latency performance doesn't dive off a cliff like it does with spinning platter drives. It instead degrades gracefully on the SSDs.

SSDs are definitely edging 15k SAS drives out of the market. SSDs do everything at 15k SAS drives can do, with at least an order of magnitude more IOPS/drive, for only about 2-4x the cost of the 15k SAS drive. And putting a writeback SSD cache in front of a spinning platter drive array is even more economical.

Re:How about transfer rate and reliability?

[Anonymous Coward]

I have about 900 CD's I've legally purchased over 20 years or more. I have one still functional CD player and a functional 6 disc CD changer. I rarely if ever use them. I also have a 3TB hard drive connected to a Segate GoFlex home. I've made legal backup rips to the hard drive(in FLAC) for about 330GB. I have several GoFlex TV media players (which play FLAC) that also support playlists created from WinAmp.

I've created playlists using WinAmp that span upwards of 50 CD's and saved it to the 3TB drive. I load the playlist to GoFlex TV and put it on shuffle and I have hours of my own music going. It becomes real handy for the holidays.

I have also ripped my DVD collection which isn't as large as my CD collection. It is mostly of kids DVD's. It is sad to say but I have atleast 20 DVD's of Dora alone.

But, what is great about the GoFlex TV is that is supports a removable portable hard drive. I take the GoFlex TV media player and plug it into the AUX ports in the Van. The GoFlex TV runs on 5 volts and I use a cigarette lighter USB charger for an iPhone to power it. I had to modify the end plug to fit into the GoFlex though, which should be easy for almost anyone.

I now have hours of movies for the kids to choose from when going on long trips with out the need to insert them one at a time in to the dash. If I don't have a passenger sitting in the passenger seat, that requires me doing it, most of the time while driving.

I also have a decent camera that I save all pictures to the 3TB drive. I've easily used up more than half the drive.

I can easily see most people using or needing more space.

Re:How about transfer rate and reliability?

[MachineShedFred]

Yeah, because there is absolutely no use for having large storage capacities with even bigger video formats on the way. Or with people wanting to not deal with optical media anymore. And there is absolutely no use in business of keeping large amounts of data online.

Have you even seen the camera density in a large retail store these days? Do you think that video might get stored somewhere for legal purposes if there's an issue, or do you think they have 50 VHS recorders in the back?

Re:How about transfer rate and reliability?

[rasmusbr]

Most of Joe Six Pack's data is going straight to his various cloud accounts, which he expects to be cheap (if not free). This means that Google and Amazon need the price per unit of storage to go down, if they want to keep their customers.

The limiting latency here is the > 50 ms latency to the servers and the limiting transfer speed is the 100 Mb/s or so offered by the ISP that Joe uses. Regular old HDD:s will do fine for the storage in that situation. The database that catalogs Joe's collection of "home movies" will probably be stored in RAM most of the time, but SSD:s do have their place here as well, of course.

This is not exactly a niche case btw.

Re:How about transfer rate and reliability?

[ultranova]

There was an article recently (can't remember where) that made the case that with slowing density increase the lifetime of HDDs has to decrease because you're not going to replace them after two or three years anyway because the next generation is so much better.

Fixed that for you. Why would hard drives be exempt from the forced obsolescence business model?

Re:How about transfer rate and reliability?

[ColdWetDog]

What does 1.701ms have to do with Star Trek?

My God.

And here on Slashdot even.

Sigh

Re:How about transfer rate and reliability?

[ColdWetDog]

MTBF is proportional to the bridge reliability.

Well this is obviously a problem.

With the Captain, First Officer and bog-knows who all jumping into the transporter for every little issue, it's no wonder that the efficiency on the bridge suffers. You're supposed to send down expendable people to the surface.

Re:How about transfer rate and reliability?

[zwarte piet]

I have my swap on a ramdisk (ducks)

Re:How about transfer rate and reliability?

[zwarte piet]

They reveal themselves when you reverse the polarity on the deflector shields.

Re:How about transfer rate and reliability?

[zwarte piet]

AND having to deal with dreadfully slow storage for that time.

Re:How about transfer rate and reliability?

[fustakrakich]

...I'd rather see drives with double the MTBF than double the capacity.

That's not the way the market sees it. Increased reliability can only hurt sales. Our economy system requires perpetual growth, or it will collapse. Planned obsolescence is a very old method of keeping the growth alive. There is no reason that a computer and all its components can't last 20-25 years like our old TVs and refrigerators did.

Re:How about transfer rate and reliability?

[Anonymous Coward]

it's the giant number painted on the Enterprise newbie.

Re:How about transfer rate and reliability?

[Rectal Prolapse]

What did you do wrong? :(

Re:How about transfer rate and reliability?

[Hamsterdan]

Not my experience (thus far).

I've had multiple 500-1TB drives fail (bad sectors and reallocations) in the last 4 years while my SSD (OCZ) is still kicking even if it spent half its life in a XP machine without any TRIM support. The spinning drives were Seagates and WD Greens.

Last summer I dismantled a whole bunch of HDDs for recycling, and you can see modern drives are cheaply built (no dessicant cartridge, less filtering and other stuff). That's the price to pay for the capacity race.

Re:How about transfer rate and reliability?

[Blaskowicz]

They could provide great catastrophic discharges.

Re:How about transfer rate and reliability?

[Blaskowicz]

And then I question that 100us figure. It's the actual current latency for accessing an SSD including the software/OS requesting the data, the driver, latency from the AHCI protocol (soon to be replaced with much lower latency NVMe), the SSD controller accessing the physical flash.

Re:How about transfer rate and reliability?

[BarbaraHudson]

But at least when a drive is getting ready to donate its' magets to the fridge door, it usually makes noises, clicks, squeals, etc. That gives time to back it up. Plus I've never heard of anyone restoring a failed SSD by sticking it in the fridge overnight (in a plastic bag, of course).

There are trade-offs either way, because unfortunately NOT failing is not an option.

Re:How about transfer rate and reliability?

[BarbaraHudson]

Much better, much faster the HDD, and no Star Trek references.

Really? You might want to re-read your post.

Re:How about transfer rate and reliability?

[BarbaraHudson]

1 TB has enough data for every man woman and child who ever lived to write a 1500 page book! Will Joe Six Pack really need to have that 30 TB drive when his circa 2012 1 TB drive has 70% free space on it?

Okay, but it's not enough to back up the consciousness of even 1 Joe Sixpack while waiting for the singularity. Anyone want to make a WAG on how much storage that would require?

Re:How about transfer rate and reliability?

[Anonymous Coward]

Even I would have a hard time finding enough porn to fill that.

That sounds like a dare!

Re:How about transfer rate and reliability?

[rrohbeck]

"Access time"="time from sending the request until data starts to flow." For 1,000 times the IOPS the transfer rate and overhead on the bus would have to scale by three orders of magnitude too, which they don't.

Re:How about transfer rate and reliability?

[Anonymous Coward]

The growth idea is insane. It basically means pyramid schemes are the only way to do things, as opposed to building a sustainable business in an industry.

You don't have to have growth to have profit. AutoCAD for example is extremely profitable, and they have been peddling the same stuff for decades.

If a HDD company made drives that had much longer usable lives, as well as longer MTBFs, they may end up with a best-seller, just because of the time and money it takes to yank out and replace drives, even on the high end VNX or NetApp platforms.

Of course, an archival grade hard disk would be revolutionary, a drive that is made to be put in a silo, is resistant to physical shocks, has multiple integrity checks for bit rot, and can last 20 years without suffering from stiction or other ailments.

Re:How about transfer rate and reliability?

[Anonymous Coward]

Dont forget to check 'amatuer' too.

Re:How about transfer rate and reliability?

[drinkypoo]

Okay, but it's not enough to back up the consciousness of even 1 Joe Sixpack while waiting for the singularity. Anyone want to make a WAG on how much storage that would require?

You mean the typical wal-mart shopper? I'm guessing 640k should be enough for anyone.

Re:How about transfer rate and reliability?

[drinkypoo]

I have my swap on a ramdisk (ducks)

don't you? I do. It's compressed, too. And what's more, you can have this on your phone. Most of the alternate Android kernels I've tried have come with zRam support, but it was also [relatively] recently added to Ubuntu as a default feature.

actual swap is so 1990s. even in the last decade I was mostly disabling it. in fact, I don't have any swap enabled on any computers, from 128MB RAM (the least of anything I've got, now — and they're pogoplug v4s and a dockstar) up to 8GB. This only caused me problems in Windows 7, where Java shits itself when trying to use 3GB out of 8GB (yes, it's 64-bit) even when I have no other foreground applications whatsoever. But then, Windows is what it is, incredibly polished in some areas, an incredible turd in others. Kind of like Linux. All I want is video from my firewire camera (which is working fine, actually, in spite of being literally one of the earliest examples, an iBot) in V4L applications. Is that really too much to ask? Apparently yes, yes it is. And sadly, it used to work, I've done it before. v4l2-loopback doesn't work and vloopback doesn't build, I haven't yet figured if the right diddling can make that happen or if the interfaces it needs are gone.

Re:How about transfer rate and reliability?

[drinkypoo]

But at least when a drive is getting ready to donate its' magets to the fridge door, it usually makes noises, clicks, squeals, etc. That gives time to back it up.

I've only had one 3.5" drive warn me with scary noises before the noises were so scary that they frightened it out of giving me any data. In fact, the first 3.5" drive I had fail did so silently but with the smell of smoke — turned out that it (a Seagate half-height RLL drive whose ST- number shall remain forgotten if I am lucky, I could use that space for phone numbers) had experienced stiction and then burned the stepper power trace off the board. By the time a jumper wire had been soldered across the trace, the drive cooled down and it spun up and functioned again. The next time it stuck hard and burned the trace off again, so I just popped the case lid off without damaging the foam gasket, reached in and spun the spindle with my thumbs and then popped the lid back on after giving it a quick puff to blow any dust off the top platter, soldered on another jumper, and it worked faithfully until I retired it. I believe that was a 40MB with at least three platters, if you could see 'em you could've about counted the cylinders. It never made a bad noise through the whole experience, and it was basically an antique by modern standards.

I did have one 2.5 inch drive develop massive bearing whine before failure, but I've had two fail silently.

Re:How about transfer rate and reliability?

[drinkypoo]

my SSD (OCZ) is still kicking even if it spent half its life in a XP machine without any TRIM support

OCZ? That's probably why it survived.

Last summer I dismantled a whole bunch of HDDs for recycling, and you can see modern drives are cheaply built (no dessicant cartridge, less filtering and other stuff). That's the price to pay for the capacity race.

So were those all consumer-level drives, or were any of them sold as "enterprise"?

Re:How about transfer rate and reliability?

[BarbaraHudson]

Darn, sorry I can't upmod you, but you gave me a laugh when I needed it.

Re:How about transfer rate and reliability?

[zwarte piet]

I tried "no swap" long ago, on a 386 with 5MB ram, but then the system would hang when the memory got full.

Re:How about transfer rate and reliability?

[drinkypoo]

I tried "no swap" long ago, on a 386 with 5MB ram, but then the system would hang when the memory got full.

long ago, you needed swap, though. And it wasn't until XP that you could reasonably attempt to run Windows NT without swap without it exploding.

Re:How about transfer rate and reliability?

[Anonymous Coward]

According to NewEgg, SSDs have about 1/2 the failure rate of regular HDs based on returns on warranty claims. That is a few years old now, and SSDs have gotten a lot better.

Re:How about transfer rate and reliability?

[zwarte piet]

Sorry, forgot to mention it was on Linux 1.2.x

Re:How about transfer rate and reliability?

[drinkypoo]

Sorry, forgot to mention it was on Linux 1.2.x

Yes, back in those days I also had a 386, although mine had 8MB RAM. I could therefore get away with only using 8MB of my 120MB ATA disk for swap. Today, I use 0MB of my 160GB SATA SSD, and also 0MB of the 320GB ATA that I use for my /opt.

Re:How about transfer rate and reliability?

[Blaskowicz]

For your issue of making your crap work, the future solution (or present solution) is probably to have hardware with IOMMU support and a VM that has your firewire controller in there. Then outdated linux can run with your firewire (hopefully!). Something like debian lenny maybe works if antiquated enough (else it was modern and piece of cake easy to install and use, like Ubuntu). I had installed a proprietary driver for TNT2 M64 on it - that made me jumps through hoops - and also xmms 1 (which in the end sucked regarding non-mp3 playback and russian / "illegal" characters) and I wonder if lenny would run on your hardware even.

Re:How about transfer rate and reliability?

[zwarte piet]

I read up on the subject and decided that reducing the swappyness to 1 rather than turning it off would be the best idea on *nix systems. Not sure what to do on Windows.

Re:How about transfer rate and reliability?

[zwarte piet]

and http://askubuntu.com/questions...

Re:How about transfer rate and reliability?

[drinkypoo]

I read up on the subject and decided that reducing the swappyness to 1 rather than turning it off would be the best idea on *nix systems. Not sure what to do on Windows.

99 times out of 100, swap is stupid. If you have lots of RAM then you never want to swap. The only time it happens is when a process goes out of control and starts allocating all your memory. Swap makes this problem worse, not better; when it happens, your system now spends all its time paging. Performance slows to nothing if swap is on HDD, and your SSD gets thrashed if that's what you're using. How many rewrite cycles does that consumer-level drive have, anyway? Buy some RAM, disable any real swap. If you don't have swap then the OOM killer cuts in earlier, and if you are lucky it kills the offending process.

Re:How about transfer rate and reliability?

[zwarte piet]

I have 2GB of ram in a P4, and usually 2.5-3GB in swap. When the swap gets over 4GB (Firefox seems very happy to do that) the system gets too slow to work with. Not sure what would happen if I turn swap off. I probably would see lots of out-of-memory errors and have to use the reset button often. On Linux I've been told, a bit of swap can actually speed up a system by freeing up ram for disk caching.

Re:How about transfer rate and reliability?

[drinkypoo]

Not sure what would happen if I turn swap off. I probably would see lots of out-of-memory errors and have to use the reset button often.

Yes, you are spot on. I imagine that you only have 2GB because your P4 takes RDRAM, which was named after the sound RAMBUS executives made while going to the bank to cash their ill-gotten gains: RDRR! Ahem. Anyway, throw that thing away! By now you can get something faster under a hundred bucks. I literally just threw away my P4 with 2GB. It's too big and consumes too much power to be worth saving. There are netbooks with more cojones.

OK, I'm privileged because I can afford to throw away a computer, even if it is a shitty old P4. But you should seriously be able to score someone's old MB+CPU+RAM for thirty or forty bucks. And by "old" I mean "way newer than your crappy P4"

Re:How about transfer rate and reliability?

[zwarte piet]

It's a Dell (with xp) and the mb only takes 2GB. I have been promised a Core-i5 machine with 8GB (refurbished Dell win win7), don't know when that arrives, but for now I'm programming with this one. Can't wait till I actually start earning money, wich should be soon (I guess that's being selfemployed for ya).

Re:How about transfer rate and reliability?

[zwarte piet]

Thanks, I enjoyed the interaction.

Re:How about transfer rate and reliability?

[Wolfrider]

--I'm with you 99% as far as ZFS. The thing is, I *don't need* a 100TB hard drive, that's a single point of failure with WAY too much data on it to try and copy off before it dies.

--I'd like to see a drive that stores multiple copies of files like ZFS natively, so in the event of $badthings, the end-user has some way to recover files. (It might be nice to have RAID1 on the same drive, with independent heads between the platters.) I'd like to see a drive that goes beyond SMART diagnostics and gives you a really good indicator that it will fail soon. And all of this needs to work well / transparently with Linux and other OS's.

--A 5-year HD lifetime should also be a *given* anymore by default, but WD Black drives are about the only ones I trust for that. (I'm open to others tho, but WD's warranty/RMA process is really good as well.) Hitachi looks promising according to the Backblaze report*, but I'm not sure about the prices and their RMA process.

* https://www.backblaze.com/blog...

Re:How about transfer rate and reliability?

[cthulhu11]

With high capacity tier 3 drives, one reason that RAID 6 (or a RAID 50 setup with tiers/groups of disks) is used is because it can take days to rebuild a blown drive.

Wish I had mod points -- this is soooo often overlooked. And one reason why I love to mirror 3 ways.

Re:How about transfer rate and reliability?

[Hamsterdan]

They mostly were older consumer drives ranging from 20 to 320GB (still had dessicants and other nice stuff).

The big surprise was when I dismantled 15k fibrechannel drives. The engineering was pretty amazing (and the magnets are really kick ass)

I should hope so!

[Anonymous Coward]

by then... we should be augmented... according to Deus Ex HR

Re:I should hope so!

[Anonymous Coward]

My vision is augmented.

First Post!

[Anonymous Coward]

I got first post bitches!

Re:First Post!

[Anonymous Coward]

vainly inaccurate distilled profane lieth thinketh sparks maintain glass
immortality circumstance inclination exult employest Sometimes distribution
restrained grass whereby recondite 1 performed tempers kings watch rebelled
contents bereaved melt joying assaying guidest knowingly literally communication
coats same service illuminate self applied driven churches usefully speedily
Furthermore reproachful bubbling impatient Duad lectured Hope trample whirlpools
publication descending Predicaments modulation severing associate shonest Print
array wheresoever Apostles invests forgetting affright vouchsafedst vaunt
condemned emblem cruelty achieve Thereupon NEGLIGENCE bared tongues predicted
cleaveth Duad it Catiline religious Inhabitant declared good period lowliness
shamelessness estate against please bounded consulted deceits Suffer copious
arrogancy bought despise prize torments fastings athirst plot factito gales
expansive feeding infer stank detached statement repress dir Despite retain
Increase worn dug detesting serves functions infants Known elements lists
exclusions perfectly covered otherwhere salted Vintage 2 mistaken Anon servitude
fleeting 1971

MO as an HDD

[evanh]

I always figured magneto-optical discs would be a good candidate for packaging in a higher precision drive. Looks like that's on it's way. :)

Re:MO as an HDD

[Anonymous Coward]

it's means it is
magneto optical??? Hey what's the #1 song there back in 1993?

Re:MO as an HDD

[ihtoit]

already been done, failed miserably.

LS120
Zip
Jaz

My favourite is the LS120 since the drives are physically compatible with the standard 1.44MB floppy. Still got an LS120 drive, it's in a short tower with my two zip drives (1 100MB and 1 250MB), my 2GB Jaz (only used twice), my 8GB Colorado DAT and my DC300 DAT. The 5.25" is in its own box since it's a half height unit and it looks silly in the tower with a hole above and below.

Re:MO as an HDD

[TheRaven64]

None of these were magneto-optical, they were all purely magnetic. Sony's minidisc was MO, and a lot of higher-end machines had MO drives (for example, the NeXT Cubes only had removable MO disks, so every user could have their own OS install and files).

Re:MO as an HDD

[ihtoit]

interesting, and I sit corrected, except that the LS120 is in fact magneto-optical. The design of the LS120 SuperDisk system came from an early 1990s project at Iomega. It is one of the last examples of floptical technology, where lasers are used to guide a magnetic head which is much smaller than those used in traditional floppy disk drives (ergo, "magneto-optical"). Iomega orphaned the project around the time they decided to release the Zip drive (which used a similar mechanism) in 1994. The idea eventually ended up at 3M, where the concept was refined and the design was licensed to established floppy drive makers Matsushita (Panasonic) and Mitsubishi. Other companies involved in the development of SuperDisk included Compaq and OR Technology. (via Wikipedia, source: Panasonic). The NeXTcube disks were MO removable hard disks, similar to Iomega Zip in the way they functioned* which is necessary for random access (otherwise you'd be sat there for hours waiting for a complete disc rewrite every time new data is committed).

*LS120, NeXT MO and Zip use lasers to guide the magnetic heads using optical tracking on the opposite side of the disc. While lasers aren't used to actually write the data, this still qualifies as magneto-optical.

**Yes, Jaz is sort of an unsealed Winchester dealy, I'll let you have that one.

Re:MO as an HDD

[m.dillon]

Jeeze, I remember those. Hey, how about the Bernoulli box? I had one of those too. Might not be M.O. though. Think it was just magnetic.

There is one basic problem with megneto-optical drives and why they've basically fallen off the edge of the earth... instead of having to have one high-precision/high-bw part in the drive you now have to have two. In the world of storage, that makes it too costly a technology to produce.

-Matt

Re:MO as an HDD

[ihtoit]

the BB was a thin (hence uber flexible) magnetic film in a cartridge IIRC. About the size of a magazine. They were touted as the big thing before optically guided magnetics came in (or even "true" MO like MD/HiMD), cheap hard drives (relatively speaking, 40MB was the size of a canned soda machine when these things hit the market) as a step up from floppies because spindle wear aside, disastrous failure was practically impossible. Head crash? Not a chance, the mechanism used a (then) 250 year old principle which states simply that the faster a gas moves between two solids the higher the static pressure, hence for a gas trapped between a nonmoving head and a spinning disc, the gas is moving, dynamic pressure decreases between the two solids (pulling the flexible disc towards the stationary head) yet with the increase in static pressure the head at a working distance of 1 micron will never actually touch the disc. Pretty much the same principle behind ground effect aircraft.

Doubt it

[Anonymous Coward]

Surely in another 11 years we'll have a newer tech and the de facto standard will be SSDs. HDD are only king for storage due to their pricing once SSDs match them for GB we'll start to see much larger capacities rendering HDDs obsolete.

Re:Doubt it

[putaro]

I remember hearing that back in the mid-80's. May I introduce you to bubble memory?

Re:Doubt it

[Anonymous Coward]

How many 640K discs would that be?
Bill would like to know.

Re: Doubt it

[Anonymous Coward]

SSDs don't look like they're going anywhere and other than price seem to be an improvement over HDDs in every other way.

I had never heard of bubble memory interesting.

Cascading

[Anonymous Coward]

We need to cascade memos about our homogenised administrative processing. Forward-looking companies invest in 21st Century management projections.The consultants recommend balanced logistical alignment.

Finally ...

[Anonymous Coward]

A hard drive to hold all my p0rn!

Slashdot promised 110 Ghz 12 years ago

[Anonymous Coward]

I'll take this prediction with a grain of salt... http://classic.slashdot.org/st...

Re:Slashdot promised 110 Ghz 12 years ago

[Anonymous Coward]

LOL it's pretty funny reading the self-assured predictions from software people. It'll be fun reading the same naive predictions about space in 10 years...

Re:Slashdot promised 110 Ghz 12 years ago

[rrohbeck]

For some reason extrapolating exponential growth is very popular in certain circles. I assume that's because it detracts from other areas of exponential growth, like total population, energy consumption and CO2 emissions. Singularity, here we come!

Re:Slashdot promised 110 Ghz 12 years ago

[Zontar The Mindless]

You talking about 110 GHz networking, or the prediction that the singularity would occur in January 2012?

Re:Slashdot promised 110 Ghz 12 years ago

[Anonymous Coward]

You mean Space Nutters? Throw in an Asteroid of Death and some drama about the species, and that sums them up.

Re:Slashdot promised 110 Ghz 12 years ago

[SuricouRaven]

The good news is that population appears to only go through a phase of exponential growth before naturally settling down. Something about a developed society really depresses birth rates. A few extreme cases have required the use of coercive population control to get through the developing phase, but even China is now in the process of gradually eliminating it as no longer nessicary.

Re:Slashdot promised 110 Ghz 12 years ago

[rrohbeck]

Not really. There are a couple of developed nations that still have a high birth rate which shows that reduced fertility is not an automatic byproduct of developmnt. The current UN forecast is 11 billion in the early 22nd century. I think mother nature will have something to say about that.

100k employees making 100k a day in email

[Anonymous Coward]

So lets say 100k employees, making 100k of emails a day each, that's 3.7 terrabytes per year, so this would hold the CIAs emails for 25 years?

So tell me again why the CIA and homeland security needs to delete all its all emails from the the early 2000s to "save space"?
http://www.engadget.com/2014/11/26/cia-homeland-security-emails/

Yet the NSA builds a new storage facility in the Petabyte class, (approaching Exobyte numbers) in Utah and has plans for 6 others!

Re:100k employees making 100k a day in email

[sexconker]

They need to delete those emails because they don't want people to see them.
Each employee won't generate 100 KB of emails per day. Deduplication and compression on the back end will shrink that massively. The more employees and emails you throw at it the more effective it becomes because we all tend to say the same shit over and over.

Re:100k employees making 100k a day in email

[tompaulco]

Because the cost of a drive to a company is somehow many hundreds of times what it is to you and me. My company measures the cost of disk at $10,000 per terabyte, whereas I can get a 3 Terabyte drive for less than $100.
My company has 60 employees and they have a 1 GB limit on e-mail accounts. I'm not sure they still manufacture drives small enough to where that is a valid limitation.
They also encourage you to back up your data, and provider a shared drive amongst all users which has a total size of 40 GB. Most of our laptops have 1 to 2 TB drives.

Re:100k employees making 100k a day in email

[WuphonsReach]

(shrugs) your IT is definitely stuck in the 2000s (i.e. 5+ years ago).

Cost per TB (raw storage, the hardware to hold the storage, plus the backup tapes / disks) for bulk storage - is definitely more like $800-$1000 per TB these days and not $10k. The sweet spot for bulk storage these days is the 3TB 3.5" enterprise SATA drives at about $230 each. Add in the loss of capacity due to RAID + server costs and you're at about $500/TB of actual storage.

Primary storage is still much more expensive at $1500-$2000 per TB. But primary storage is using SSDs (around $1/GB) or 15k SAS drives (about $0.35/GB to $0.50/GB). And not the relatively inexpensive 3TB enterprise drives at $0.08/GB.

More than Commander Data's Memory Capacity

[Anonymous Coward]

He states he has a memory capacity of eight hundred quadrillion bits (TNG: Measure of a Man).
(((((800x10^15 bits) / 8bytes) / 1024kb) / 1024mb) / 1024gb) / 1024/tb) = 90949.47 TB

Re:More than Commander Data's Memory Capacity

[Brad1138]

I am sorry, when is 100 TB more than 90949 TB?

Re:More than Commander Data's Memory Capacity

[Anonymous Coward]

When I post as AC :D

Re:More than Commander Data's Memory Capacity

[ihtoit]

Johnny Five had a storage capacity of 400MB. More than enough to digest the entire contents of a library and still wanting more. Yes, even with that, they somehow managed to add a lightning strike, simmer and add emotional responses and have space to spare. I can't find a decent OS with voice recognition out of the box (never mind a heuristic analysis and response) that takes less than 20GB in the initial install!

Re:More than Commander Data's Memory Capacity

[tompaulco]

400 MB is enough to store a library. The problem is, a picture is worth a thousand words, but it takes the space of 100,000. And a video takes the space of 10 or 20 pictures per second of video.
Pictures and video is what a lot of the large drives are being used to store.

Will man still be alive?

[JeffElkins]

https://www.youtube.com/watch?...

It wont be that long.

[Brad1138]

I remember 1GB drive going for ~$100 in 1998. 1TB drives have been under $100 for a few years. We will probably see 1PB drives by 2025

Re:It wont be that long.

[daniel23]

I remember when I upgraded from my Armstrad cpc 6128 to a 386sx machine which was my first box with a HDD. I choose the large one then with 40MB. Somewhen at the last third of the eighties. Those things were expensive then and I could have got a used car for the price of it

So that means we're still gonna be buying

[Spy Handler]

2 types of drives for the foreseeable future.... SSD main drive for the OS and programs, HDD for storing pron. Better get used to it.

Here is my desktop setup:
- One fast SSD main drive
- One 3 TB HDD for video/picture/mp3/document storage
- Second 3TB HDD, gets synched to the first HDD nightly as a backup

SSD gets a manual disk image stored on the first HDD once in a while.

I feel pretty good with this setup, as I'm protected from any single drive failure. Also I have some accidental deletion/corruption protection as the two HDDs are not mirrored in real time.

Re:So that means we're still gonna be buying

[Billly Gates]

The question is will your 3 TB ever get filled? That is a lot of storage.

Also dropbox and MS Onedrive built into Windows 8.1 both offer 1 TB of cloud storage too. This will decrease demand for consumers.

We are nerds and not typical users but even I have a similar setup ... well with a cloud for extra backup in my case. I have 1 TB for my profiles and backup data. 2 ssds in raid 0 for OS and virtual machines and some games. Increasing storage space for a mechanical disk to me is like trying to sell me yet a larger tape drive.

My prediction is hard disks will turn into tape drives of today for backups and long term storage. SSDs are the wave of the future. My prediction also is many Walmart pc wil use disks for years to come due to OEMs being greedy about costs and not value. Joes who do not know what an SSD is will like the bigger storage in the specs and cheaper price not knowing it will significantly slow him down. The same ones who bought Pentium IVs because they had higher MHz a decade ago thinking they were better.

Re:So that means we're still gonna be buying

[fnj]

OK I'll match your anecdotes.

It's nice for you guys that your needs do not encompass any significant amount of data, but there are plenty of guys with a lot more imagination and requirements. I have a personal server with two 6-drive (3 TB each) RAID-Z2 double-redundant pools. That's a total of 24 TB usable storage.

The two pools are currently 85% and 78% full.

My intention is to have a second server with the same storage specs as the first, with the pools synced to those in the first server. That will give me four-level redundancy. So far my second only has one of the pools installed (I ran out of funds).

If I scare up the funds, you can damn well bet I will not only finish the second server, but I will upgrade the disks to at least 6-8 TB each, and hopefully more.

Oh I also have 6 more 3 TB drives (so 24 total) and 22 2 TB drives installed in various other boxes. Let's not even count the pile of 1 TB and smaller that I don't even consider worthwhile to have installed any more.

Re:So that means we're still gonna be buying

[Anonymous Coward]

Try running CFD calculations on a small cluster. I fill up 3TB in a couple of days.

Re:So that means we're still gonna be buying

[Anonymous Coward]

You can just torrent that porn again if you lose it.

Re:So that means we're still gonna be buying

[ihtoit]

great.

Until you have a flood.

You still have a single point of failure, no matter how exotic your wiring.

Re:So that means we're still gonna be buying

[Anonymous Coward]

That is my exact setup except the first HD is synced to my nas.

Re:So that means we're still gonna be buying

[m.dillon]

Every personal laptop/desktop/server box has a SSD which holds the machine base (boot, swap, root, home dirs, etc). For laptops and workstations, the SSD also holds nominal data and there is no HDD at all. On the servers, the SSD is beefier and also caches the HDD.

Bulk data is stored on a large 2TB HDD on the server and exported via NFS and samba. There's another 2TB spare on the server.

There is a backup machine on the LAN in another part of the house with a 2TB HDD and there is an off-site backup machine in a colo with two 2TB HDDs (in addition to the normal SSD as described above). Both have around 60 days worth of incremental backups of everything.

I gave up using RAID of any sort for personal data long ago. It just makes things *less* dependable and unnecessarily expensive in heat and power. Not even needed for speed since the SSD will cache ~200GB worth of HDD data.

Right now my personal data (the permanent data that I back up) clocks in at around 1TB. 90% of it is pictures and videos (DSLRs generate a godawful amount of data, the RAWs are 20-40MB each).

I don't download videos or music. No point any more when it can all be streamed. I don't know anyone who bothers storing 3rd party video any more. However, I have friends who were CD junkies before CDs died away and do maintain large music libraries. I also know a few people who run torrents and sometimes dedicate a TB or two to that, but certainly there is no reason to back up something like a torrent let alone waste a RAID on it. I don't see the point myself, it's just a huge waste of power and bandwidth.

For the DragonFlyBSD project we run a 12-blade server in the colo, each with four 2.5" drives. One or two SSDs (boot + swap + hddcache + home dirs on most of the blades), and 1-3 HDDs for temporary bulk data which we don't bother backing up. e.g. build boxes. With a few spare blades in case something fails. SSDs hold anything important, except for the developer blade but now that I look at it, the 'backed up' portion of peoples home dirs on the developer box only clocks in at 83G so that could go onto the SSD as well. The working storage that isn't backed up is currently running ~400G or so of used space, mostly crash dumps and copies of build trees and such. Everything that we care about is backed up locally and remotely but again it only amounts to a ~1TB or so. Most of the bulk data on the blades, like copies of numerous source repos, is generated and does not need to be backed up.

-Matt

That's a lot of

[Anonymous Coward]

Porn

Porn

[An0nymous Coward]

100 TB is an incredible amount of porn! All slashdotters will want this!

Re: The Cloud

[Anonymous Coward]

Well, these can make those cloud servers much smaller with far fewer drives needed for such high capacity storage.

Limitless?

[hcs_$reboot]

In a few more years, everybody will have enough data space in a laptop to store the whole humanity data (as long as it's not recursive). The Internet in a cache...

Re:Limitless?

[Anonymous Coward]

What would the Internet be backed up to? A gazillion floppies?

Re:Limitless?

[toxique]

still not enough to store all my pr0n

heat and patterned-media

[markhahn]

interesting that these density improvements could both be applied to tape as well.

yeah, "tape yuck", but it makes a certain amount of sense for cool data. which we have lots of, always increasing. tape seeks are a minute or so, and if density is competitive, tape has a good chance to beat disks on price. certainly on power. the real problem is that the tape industry seems to be sort of demographically challenged...

So what happened to shingled?

[Tough Love]

So what happened to shingled, is it dead?

Re:So what happened to shingled?

[Anonymous Coward]

Nah, a lot of people get singles when they get old. There is a vaccine I think.

640K should be enough for anybody, right?

[King_TJ]

I think you're neglecting the fact that with larger storage capacities come new options.

Sure, right now you see a lot of people with a 1TB drive that's not nearly full. But I also see quite a few people who fill up really large amounts of drive space with photo, music and video libraries. With enough cheap storage, more people can store music in an uncompressed, lossless format (like FLAC) instead of compromising sound quality with MP3 or AAC just to save space. Digital cameras have gone from 1 or 2 megapixel to 14-18 megapixel in many cases, generating much larger image files for still photos, too. And even your commercial video games are taking up exponentially more drive space than they used to, as developers decide to tell stories with full-screen hi-res video, vs. scrolling a few lines of text up a screen to summarize things, and as they build large 3D worlds you can run around in and fully explore.

Another culprit for sucking up disk space in a corporate setting is DropBox. Companies using the paid "Teams" version wind up with everyone's copy of the software downloading and syncing ALL of the content stored in the shared folders. So instead of just keeping YOUR data, you now have a copy of the whole team/division/company's data you're all storing there. (Yes, you can use "selective sync" to trim this back down. But except for folders you know you have no use for at all, it's preferable to sync it all so you have immediate access to anything your co-workers intended to make available, even if you're not by an Internet connection.)

Re:640K should be enough for anybody, right?

[Anonymous Coward]

I'm one of those people. In addition to my stuff, TV and movies for 2 other people are stored on my system because I'm the only one with a proper desktop/NAS setup. I've got about 275 HD movies (4GB-40GB, mean 12GB), 180 more in SD, and 94 full tv series.. a rough estimate on that is maybe 60 seasons in HD, but a lot is old and only available in SD (I'm looking at you B5). I like having the highest quality copies available... so once 4K comes along relying on my current mashup of storage won't cut it. As it is, I have 1x4TB, 1x2TB, 2x1TB, 5x500GB NAS as individual disks- 10.5TB coming in at 94% capacity.
Could easily fill up 100TB once 4k becomes popular. And have cheap reasonable drives to back up on. Right now I can't back up anything outside of critical documents, source code and photos (5gb for everything) and have to re-DL everything on a drive failure. Would be nice to be able to afford more drives, but the point is if capacity outpaces media storage requirements it would drive down the cost of being able to actually have backup drives for us poor folks.

Re:640K should be enough for anybody, right?

[deroby]

Do I understand this correctly ? You're willing to fork out quite a bit of money to have 'a proper desktop/NAS setup'; you 'like having the highest quality copies available' but when it comes down to it you'll stick to "(re-)DL everything".

Can't say I blame the producers for not investing in a HD version of B5.

trillions of bits, why one head per platter?

[raymorris]

Somewhat off topic, but while we're talking about drives:

We put millions of transistors on a chip. Millions of photodetectors (pixels) in your phone's camera, a million pixels on it's display. Yet our hard drives have ONE sensor that swings back and forth on a mechanical arm?!?! Why the heck isn't the read/write head a strip, with a few thousand "pixels", so it can read any sector as the platter spins beneath it, without swinging the heads back and forth? That would eliminate seek time.

  If needed, you could move the strip back and forth a thousandth of an inch to align a head with one of it's four tracks. That'd be a lot quicker that moving the head a full inch as they do now.

So presumably there is some good reason that can't be done. Still, an additional arm exactly like the existing one, but on the opposite side of the platter, would cut rotational latency in half and increase throughout up to 100%. Seems like an easy win.

Re:trillions of bits, why one head per platter?

[Anonymous Coward]

...and yet, you can't figure out that it's means it is.

Re:trillions of bits, why one head per platter?

[Anonymous Coward]

Before hard disk drives, they had magnetic drums. It was a large drum that spun at a constant speed and had a separate head per track. You could read from any track, but only once the data you wanted to read rotated to be under the head for that track. Once upon a time computers used drums for main memory and core for cache.

Your idea for an arm with a head for every track is essentially a more modern drum. That and the two-armed HD are both ideas that are likely more expensive to be worth it. The head mechanism is the expensive part of the drive, so having two of them is almost as expensive as having two whole drives, so why not just have two drives?

dom

Re:trillions of bits, why one head per platter?

[amalcolm]

Most disk have multiple platters, so already have multiple heads simultaneously reading multiple tracks. Replicating the actuator (motor+magnet + electronics) would be the expensive bit.

Re:trillions of bits, why one head per platter?

[drinkypoo]

an additional arm exactly like the existing one, but on the opposite side of the platter, would cut rotational latency in half and increase throughout up to 100%. Seems like an easy win.

Drives like this exist, I don't know if any are being sold at the moment but they've certainly been made. I believe the strategy has been used both for dual-attach and for increasing throughput. Some drives also used to read/write multiple tracks (that is, on multiple platters) at a time, but (as has been covered elsewhere in this thread) it got to be too complicated to keep everything aligned as the temperatures and rotational velocities increased.

Re:trillions of bits, why one head per platter?

[tlhIngan]

Drives like this exist, I don't know if any are being sold at the moment but they've certainly been made. I believe the strategy has been used both for dual-attach and for increasing throughput. Some drives also used to read/write multiple tracks (that is, on multiple platters) at a time, but (as has been covered elsewhere in this thread) it got to be too complicated to keep everything aligned as the temperatures and rotational velocities increased.

Alignment isn't an issue - there's no alignment on a modern drive. Instead, at the factory, they write a set of servo tracks all over the platters which do the aligning for you - basically the head seeks to approximately the right position and starts reading, and the servo track tells it where it actually is, so feedback gets the head to the right track.

Old hard drives were open-loop - you said seek to track N, it went to track N by using a stepper motor. Modern hard drives are closed loop in that they are constantly looking for servo tracks to tell them where the head is to pinpoint the right track. It also allows for individual platters to have their own servo tracks before being assembled, as well as handling thermal expansion dynamically (old hard drives would do a scan every 30 seconds or so to find out where the tracks were - called "thermal recalibration". Modern ones don't need this since feedback automatically gets them there, which has the advantage that drive accesses are not paused during the recalibration). This was important if you were streaming data to or from the drive, which was why the early ones were called "A/V" drives - they were designed to do the recalibration on the fly so you can constantly write or read data.

No, the bigger reason why two actuators didn't work is far simpler - think multiprocess programming. Both actuators could read or write data to the platters (of which there was one set) and if you screwed up the order of the accesses, you could easily write the wrong thing (think you do a read then a write of a sector - and the sector happens to be under the actuator doing the write). And yet, if you serialized the accesses, you're back to square 1. So maintaining data consistency was incredibly difficult and at higher datarates, unmanageable.

Re:trillions of bits, why one head per platter?

[drinkypoo]

Alignment isn't an issue - there's no alignment on a modern drive. Instead, at the factory, they write a set of servo tracks all over the platters which do the aligning for you - basically the head seeks to approximately the right position and starts reading, and the servo track tells it where it actually is, so feedback gets the head to the right track.

Sigh. Alignment is an issue, because each platter has its own alignment. That means that when you're reading/writing one platter, you're not aligned for the other platters. That's why you can't have multiple heads on one armature (which has multiple arms, all fixed together) and read/write multiple platters at once.

the bigger reason why two actuators didn't work is far simpler - think multiprocess programming. Both actuators could read or write data to the platters (of which there was one set) and if you screwed up the order of the accesses, you could easily write the wrong thing

You're being ridiculous. That's true no matter how many actuators you have — if you screw up, you write the wrong thing. Even if you only have one actuator, if you write the data to the wrong sectors, you're gonna have a bad time. But both actuators have the same job: write some data to someplace. The two don't have the job to write the same data. If the drive gets a command to write data to a sector to which it already has cached data waiting to write, then hopefully it just throws away the first command anyway. This is something we would hope any drive with queuing would do whether it has 1 actuator or a dozen.

think you do a read then a write of a sector - and the sector happens to be under the actuator doing the write

HDD sectors are either 512 bytes or 4kb. In the former case they are often smaller than filesystem blocks and there is no need to read them before writing. You just run right over them. In the latter case, they are typically the same size as filesystem blocks (we use bigger blocks on larger filesystems, and we use 4k blocks on multi-TB drives) and again, there is no need to read them before wrtiting. You only have to find them, which means waiting the seek and then for some fraction of the time it takes the spindle to go around once. Then you can write. This is true no matter how many armatures are reading/writing the same disk.

Re:trillions of bits, why one head per platter?

[toddestan]

In every drive I've ever taken apart, all the heads move together so you're not reading data simultaneously from multiple platters unless by chance you just happen to need to read data from the exact same spot on multiple platters at the same time. Even if this happened, my guess is that the drive isn't capable of utilizing more than one head at a time anyway.

SSDs and Flux Capacitors

[Anonymous Coward]

Will SSDs with bullt-in flux capacitors be available by 2025?

Cheap laptops

[jones_supa]

Cheap $300 laptops these days slap in a 500GB HDD to satisfy the requirement of a hard drive. A basic 500GB disk is what they can source cheaply and easily from the market. However, I suspect that they could ship a smaller capacity disk as well, if that allows the manufacturer to shave off some of the laptop's price. What follows is, that I also suspect that when making a 128 GB SSD becomes cheaper to manufacture than a mechanical HDD, many low-end laptops will move to the SSD format.

Re:Cheap laptops

[Blaskowicz]

Happens sort of in announced $200 laptops. You get 32GB eMMC soldered on the motherboard and can't add or change anything - beware of what you ask for lol.

Re:Cheap laptops

[drinkypoo]

I haven't laptop-shopped in a long while, I'm kind of awash in them right now, but last time I looked even most fairly low-end laptops were offered either with a small SSD or a larger HDD; say, 40/250, 80/400, 120/500, something like that. The very-lowest-end machines (netbooks) were coming with as little as 4GB flash, but up to 16 or rarely 32GB as you say. It was however often on a module that you could upgrade if it wasn't already a 32GB.

Re:Cheap laptops

[Blaskowicz]

Indeed, those were some weird ass PCIe SSD. Had to install linux on a friend's one which had 4GB + 8GB and the memory upgraded to 2GB, very good survival could be achieved ( / on ext2 on the 4GB, /home on 8GB, no swap )

The new stuff coming is a bit different as they're a System-on-Chip instead of a CPU + chipset, so they make it more integrated on a very small motherboard and everything is soldered like on a tablet (or those stupid macbook pro and 2014 macmini). Hoping a manufacturer with basic respect for human feelings and the underclass will make one but with an additional SO-DIMM slot and M.2 or mSATA.

not a lot of use for most

[ihtoit]

...especially when the *AA congloms get their way as they usually do in forcing ISPs to block certain content.

"But it's to protect the children!" Bullshit, try shutting down the child traffickers accounts on facebook and ban advertising for foster carers for financial incentives - in fact, ban financial incentives for looking after other peoples' kids and instead try helping the families instead of making shit up about them. The best place for a child is with the family he was born into, NO EXCEPTIONS. If his entire family is dead, THEN you can talk about adoption, otherwise it's not adoption, it's trafficking.

"But it costs the artists money!" Bullshit, musicians don't make anything on CD sales (the last person who did died of a drug overdose, his name was Michael Jackson and he sold millions of records - not even the Beatles made money selling records until after they split up and Lennon got his face shot off). Musicians make money on concert footfall, and then they're paying royalties to their LABELS. Actors and filmmakers are paid in advance of publication, and their contracts stipulate no royalties - only the producer gets any royalties since it was he who stumped up the capital for the project in the first place!

Re:not a lot of use for most

[Kjella]

"But it's to protect the children!" Bullshit, try shutting down the child traffickers accounts on facebook and ban advertising for foster carers for financial incentives - in fact, ban financial incentives for looking after other peoples' kids and instead try helping the families instead of making shit up about them. The best place for a child is with the family he was born into, NO EXCEPTIONS. If his entire family is dead, THEN you can talk about adoption, otherwise it's not adoption, it's trafficking.

No exceptions? Have you any idea how bad violence, sexual abuse, neglect and so on can get? In the worst cases, children die. Foster parents are paid because they tend to get deeply traumatized kids with behavioral problems who need lots of care and therapy. The alternative is often institutions because leaving them with the parents was not an option. You rarely if ever get paid for adopting healthy, normal children because there's many childless couples that'll do that job for free.

Re:not a lot of use for most

[ihtoit]

come back to me when you've had experience in public family law.

Re:not a lot of use for most

[Anonymous Coward]

The best place for a child is with the family he was born into, NO EXCEPTIONS.

And your omniscience, O God, lets you make this universal pronouncement, yet does not reveal to you that we ordinary mortals recognize such unqualified generalizations as the sign of a weak and simple mind?

Stupid absolutist fucks are stupid.

Re:not a lot of use for most

[drinkypoo]

The best place for a child is with the family he was born into, NO EXCEPTIONS. [...] musicians don't make anything on CD sales (the last person who did died of a drug overdose, his name was Michael Jackson

And his father beat the living shit out of him. And look at how well he turned out! Boy, I'm sure he's happy today!

Re:not a lot of use for most

[ihtoit]

I'm pretty sure he's dead.

Re:not a lot of use for most

[drinkypoo]

I'm pretty sure he's dead.

So you can agree that being raised by his own parents didn't work out so well, right? He destroyed his face out of low self-esteem in spite of being one of the best-loved entertainers in history, and died of a prescription drug overdose. Now, can you prove that being raised by someone else wouldn't have been better for him?

Re:not a lot of use for most

[ihtoit]

no I don't agree on that at all. Being brought up by someone else: can you prove it could? I can provide ample anecdotal evidence that being brought up by anyone other than natural parents breeds psychopaths (R V Stafford 2010 where a 24yo out of the care system and just out of prison on parole for arson reckless immediately attempted to murder three entire families, including his own, by petrol bombing their homes in the middle of the night: R V Bacon (1994?) 18yo just aged out of the care system jailed for six years for raping a baby. Did either of them know any better? I think not, beyond feeding and clothing kids in care aren't given a moral compass, in fact their so called foster carers don't give a fuck about them beyond keeping them alive to collect the £400 per child per week Government stipend. Social workers will lie to foster carers, and foster carers will lie to charges. I have documented evidence of this, there is also ample evidence in the public record). Oh, by the way, MJ didn't kill himself, his doctor killed him - this is also a matter of public record (People of the State of California vs. Conrad Robert Murray Case Number: SA073164).

Re:not a lot of use for most

[drinkypoo]

I can provide ample anecdotal evidence that being brought up by anyone other than natural parents breeds psychopaths

Oh, so nothing substantive then. Just as I believed.

Oh, by the way, MJ didn't kill himself, his doctor killed him

Yes, that's quite common.

Re:not a lot of use for most

[ihtoit]

I gave you two very substantive examples out of many. BAILII is bursting with them.

Re:not a lot of use for most

[drinkypoo]

I gave you two very substantive examples

There is no such thing, only substantive evidence, which test your examples fail. You've utterly failed to show anything whatsoever.

Re:not a lot of use for most

[ihtoit]

oh fuck off.

Brave new World

[Anonymous Coward]

At that density, hard disk drives could conceivably hold up to 100TB of data.

100TB, that is a LOT of porn.

it`s a brave new world we live in, a brave new world

Who would need that ?

[bobjr94]

The average new computer being sold today sits on a desk in an office or is a facebook machine at home. It will have in it 500GB to 1TB of storage on average ? Then whats the actual usage, 20% or less ? I have people asking me all the time, my computer is slow do I have too many pictures on it ? I look at their drive, 482GB capacity, 404GB free.

Sure there are some users who have hundreds of movies stored on their computer and businesses and datacenters who would love a drive like that but by number of computers, thats a small percentage. A majority of computers would do much better with just a SSD 1/4 the size of the HDD they currently have. A faster system overall, bootup times cut by 60%, 20-30 minutes more battery life in laptops....

Re:Who would need that ?

[Anonymous Coward]

Um.. No, I got this.. Um.. Data-Centers and the NSA? Plus Harddrives that large could potentially replace LTO tape back-up.

Re:Who would need that ?

[comrade1]

Yeah, ok. Because your grandma doesn't need 100TB the rest of the world doesn't either...

Re:Who would need that ?

[drinkypoo]

Yeah, ok. Because your grandma doesn't need 100TB the rest of the world doesn't either...

I think they're right. Most users would be better served with the SSD, because most of them aren't filling up their drives and most of those that are could probably throw away a bunch of crusty old data without remorse. The claim was not made that nobody needs the storage space, only that few need it. In my experience, it's probably true, even though I'm someone who has 1 and 3 TB disks online, and backup disks for them. That's not particularly much, either, I just know that I'm a niche case. Most people just buy a PC that costs however many dollars they want to spend on one, plug it in like it says, and open the web browser.

So much storage for what purpose?

[mark_reh]

We'll all be able to store all our "home movies" and photos on one drive!

Re:So much storage for what purpose?

[Anonymous Coward]

Never underestimate Adobe and Microsoft's ability to produce bloatware.

Kryder rate

[LeadSongDog]

To increase tenfold in 11 years, the Kryder rate https://en.wikipedia.org/wiki/... would have to jump from 15%/year to 23%/year. While this is not fundamentally impossible, in an era of diminishing revenue for magnetic storage I can't see it happening.

Who needs HDD anymore

[ssufficool]

I boot from the cloud.

Cache and NCQ already this

[raymorris]

This cache and especially native command queuing (ncq), the drive ALREADY has to pay attention to the sequence in which operations are carried out. A read requested first, then a write, might already be done in reverse order, requiring a check that the sector read isn't the same one written.

I don't see any reason reading and writing two sectors at a time makes any fundamental difference.

640K should be enough for anybody, right?

[Anonymous Coward]

Another culprit for sucking up disk space in a corporate setting is DropBox. Companies using the paid "Teams" version wind up with everyone's copy of the software downloading and syncing ALL of the content stored in the shared folders.

Why do anyone bother using dropbox like this? In a 'corporate setting' no less? Never heard of "file servers?"

Dropbox can be useful for sharing some files with family or friends that can't use sshfs or a webserver of their own. But a corporation should be able to set up a simple file server, no need to store the same shit on everybody's computers. (And if they need access out of office - a VPN will do that in a safer manner than dropbox.)

Big disks shift the balance

[Anonymous Coward]

Even 20 TB disks in a desktop will shift the balance of data storage from the cloud to the disk. The cloud becomes backup, and that brings up another issue...efficient and reliable restore. Backup we got. Restore at even 1 TB? GOTCHA.

Re:It Cannot Be Disputed

[Anonymous Coward]

This crap read like lorem ipsum.