Watch out, PC disk drive floggers: Cloud will rust up those spinners The cloud is going to impact disk drive manufacturers' bottom line, big time. Think about it: if Dropbox, iCloud and SkyDrive succeed, people won't need disk drives on their PCs. Are you listening Seagate, Toshiba and Western Digital? The biggest ever threat to their consumer computer disk drive business is cloud-based file 'n …
If you don't have a reliable ISP service then there will be times when your cloud will be out of reach. I'm with Virgin Media on a fairly new estate laid out with cable. The service has moments were it is excellent, very slow, very very slow and not working at all. The service is not consistent whereas a local disk drive tends to give to give you the data as and when you need it.
Andy
And a fricking huge wad of cash. I think the authors point is valid for people in his situation, but not everyone. I don't have a reliable net connection and I have approx 20TB of video and pictures (all legit before you comment, ex pro photographer) and this grows at about 3TB a year. Right now it costs me about 200 bucks a year in external drives and about 40 a year for a bank deposit box to put a copy in. I haven't a clue how much that would cost for cloud storage but it would take a while to upload!
The cloud will work for some (those without huge amounts to store, with reliable connections everywhere they go, and with multiple devices), and not for others (those of us with lots to store and piss poor connections).
Stating that the cloud is the end of domestic hdd sales is highly amusing. I don't eat at mc donalds so they must be going out of business!
Sorry, I don't buy it. I use Dropbox and the major limitation on its use is that my Notebook hard disk is so small I can't keep everything I want inside Dropbox.
Dropbox is a replicator not a network drive. I wish I could tell Dropbox I don't want THAT folder on THAT device. (You could simulate the effect by having multiple logons and sharing folders between accounts but I've got one paid for account and don't see why I should have to have multiple paid accounts to make up for a product deficiency)
Time Machine also sells disks. When my Time Machine cheap USB drive is full, I label it, put it on the shelf and get another from Maplins. I've got 5 on the shelf.
Also just bought a pair of 1TB cheapo USB drives for a media server. Wishing I'd bought the 2TB ones now. Of course I can add another pair of 1TB drives but that's 4 mains sockets (and power supplies) just for the disks.
We are a Very Long Way from being able to use The Cloud as the one and only storage medium. For backup it has its uses but as a primary storage medium, no way Jose. How did it go with Chrome OS, again?
Just remembering when Office Automation meant the advent of the paperless office. As I recall the main beneficiaries of that were HP, Epson, Canon.
'The cloud' certainly will have an impact on drive manufacturer's bottom lines - instead of your data being on one cheap consumer disk, it will be on 2 or 3 or 4 more expensive quasi-enterprise level disks.
It isn't actually a literal cloud you know, where bits just float around in space.
1) Your data is suddenly slower to access and update.
2) Your data is now reliant on some the 2nd party to ensure it is available.
3) With a disk, you pay your upfront cost, and then it is yours. With cloud, you pay once and then every month.
4) Your data is now at risk of some 3rd party removing access to your data. Your suppliers' ISPs, carriers, your own ISP, and potentially even hostile governments can make your data disappear.
5) Online capacities are minuscule compared to offline capacities.
Agree with Tom 38 all points, particularly point 1. Access to the cloud, especially through mobile data, is excruciatingly slow. Just "syncing" a single 8MP photo can take forever.
Having your own cloud - like Pogo Plug or your own home brew - is a partial answer, solving points 2,3 and 4. Also cheap on electricity if the backing storage is solid state. Still slow for your mobile devices though, and if the pogo plug is burgled, so is your data.
I just checked and there were several thousand copies (seeders) of e.g. the latest episode of some TV show called futura* available on some website sharing bittorrent links.
Imagine instead how much disk space would be saved if instead there were some central location on the internet storing just one copy of the file, with fast access for everyone.
Maybe this would be bad for the storage manufacturers; I, for instance might not see the need to purchase more storage so soon.
Imagine instead how much disk space would be saved if instead there were some central location on the internet storing just one copy of the file, with fast access for everyone.
Try imagining this: imagine that instead of 1,000's of locations each with full copies or one location with a single copy, there was some sort of "middle way" where we have 1,000's of locations that each store a cryptographically-produced "share" of the file such that collecting a given number (say 50) of the shares is enough to reconstruct it at the download location (with any fewer giving you zilch). This sort of algorithm actually exists, and is somewhat related to RAID-style redundancy. See Shamir's "Secret Sharing Scheme" for the details (and one implementation, or look up Information Dispersal Algorithm for another).
Comparing the secret sharing with storing 10,000 copies, a secret sharing scheme with 10,000 sites and 50 shares required to reconstruct the "secret" would only take up 1/50th of the overall space that full copies would, as each server would only by storing shares that are 1/50th of the full file size. And yet, 10,000 - 50 sites can fail and the data can still be recovered...
Some archival systems actually use such schemes behind the scenes instead of using regular RAID systems. It might be interesting to see if it would be a viable alternative to Bittorrent...
6) Bandwidth caps from ISPs.
I don't know what it's like in the UK, but in the US, even home internet isn't truly unlimited anymore - and the trend is for more limits, not less. Even really high sounding caps can evaporate quickly if you're watching a lot of movies that you're storing in the cloud...
I don't use any cloud services and, as long as I have a choice, I never shall - mainly because I don't / can't trust people with my stuff, but practically because I wouldn't have access to it most of the time. The concept might be great if you live and work in central London or Los Angeles - where WiFi can be sniffed out and 3G coverage is good, but my experiences in most of the UK is that I'm, effectively, disconnected from the web most of the time. I need (and have) two phones so I have simultaneous Vod and O2 just to guarantee voice calls to clients as I roam about the UK. Last week (Mon-Fri) I only had 3G access for about 4 hours (Norfolk, Scotland) - and a lot of the time there was no mobile signal at all. Most of my written comms is done by SMS because email isn't really available. I dream of Edge!
I'm sure that I'm a tail-end charlie on this and the article is more-or-less on the button and I guess that in ten years' time I'll be paying a fortune on eBay for those old drives (disk or solid) - a bit like the money I still spend today on Vinyl ;-)
"Uh, how about using the cloud as (an additional backup) for the stuff one finds important." - -precisely because it's important stuff. No point having important stuff backed up to somewhere I can hardly access and which could disappear at any moment.
I do daily backups when I'm at home. I carry a backup disc with me when I'm on the road and keep it up to date (mostly). Additionally all my "working" docs are backed up in plain on an Iron Key so if my PC dies a big death on the road I can be up and running with clients pretty quickly (albeit after the expense of a new PC).
There used to be a scam where a virus locked all the files on your PC and they were only released after a payment. The way that the "Cloud" is developing feels a bit like a legalized version of this. I'm sure there are some people for whom having files current on many devices via the cloud is useful and I think they're being sucked into finding it indispensable before they suddenly start getting hit for ££ to maintain a service they used to get for nothing. I'm not one of them and should the day come that I need global access to all my stuff across more than one device I'll pay my ISP the extra for a business account and static DNS and stick it all on a server.
" The cloud is going to impact disk drive manufacturers' bottom line, big time. Think about it: if Dropbox, iCloud and SkyDrive succeed, people won't need disk drives on their PCs. Are you listening Seagate, Toshiba and Western Digital?"
Have you actually USED any of these? All the content is mirrored/copied from your local hard disk, you need the same space locally on at least one machine you own as you do in the servers - i.e. the 54GB I currently store within the Dropbox folder on my media PC is mirrored to the cloud, not moved to the cloud. None of these cloud services are workable without offline copies or do much to reduce the GB needed as local storage.
If anything hard drive manufacturers should win twice, sold a drive to the consumer AND sold a drive to the cloud host mirroring the files.
However you should notice from your scenario what will reduce spinning disk shipments is not cloud services at all but the usage of alternative drive technology, something you actually state as you now have two laptops using Flash SSDs! If Western Digital and Seagate hope to continue their market strength they need to find a technology to compete on speed and reliability with NAND flash drives and then start making them at a good price.
That said alternative drive technologies are still an order of magnitude pricier per GB and less durable on repeated writes than good old spinning disks - the hard disk manufacturers are selling bucket loads of high margin enterprise qualified drives (still miles cheaper per GB than enterprise Flash SSD) to these cloud hosts. The spinny disk makers have some time to find an alternative faster drive tech to catch up with, it's not as if Flash is perfect.
I'm probably not a typical user, but what I'm mainly using Dropbox for is to provide a synch mechanism between my real XP installation and the version that I run in a virtual machine under Linux. Even then, there are actually only two applications that I regularly run in the VM that I want to keep synched, and the amount of data isn't very much at all.
I wouldn't trust these kinds of services for backups unless (a) I had some sort of front-end encryption meaning they couldn't snoop on what I'm storing and (b) I already had better/more secure backs in place anyway (say using something like Dropbox for daily backups, but making sure I do my own weekly one to my own backup box). They are kind of nice to have for "ad-hoc" backups as you put it. I could see myself using the "selective sync" option in Dropbox to set up a spare machine as an occasional repository for (relatively small) backups. I like the way that you don't need to power up the machine straight away (just get to it when it's convenient and start Dropbox to sync into your local copy) but obviously it does mean double the data transfer burden on your net connection. That's not an issue if you're syncing between home and work machines, though, and one or the other is firewalled so you can't do push transfers.
So actually I think that it's making data transfers and file sync easier is the main advantage of these things--and not actually as a primary/backup source for your data. Also remember not to trust that they won't read your stuff or screw up your files irrevocably once in a while.... otherwise, it's a pretty nifty tool.
As the recent years have gone on, the all you can eat bandwidth deals have gone. Mobile data deals still aren't up to the job of shifting the kind of data that I chew in my desktops.
As I see it, the biggest threat to hard drives over five years is whether flash drives will replace them, (assuming that our data requirements don't grow ahead of their capacities; my requirement at home is currenlty 5TB) but flash drives can fail just as easily and in just as dramatic a style.
The limitation of bandwidth allowances will most likely keep cloud services where they are; as a means of synchronising relatively small amounts of data between devices and sharing odd files. Also, who wants to not be able to work on that important letter because of a failure somewhere else in the system.
How the ruddy hell am I supposed to shove the 800gb I currently have sat on my drives on "the cloud" (or as we used to call it before marketing pricks cam along, a "hosted soloution" when I have a 1mb upload speed?
If you haven't noticed we are using more and more drive space these days, not less.
Cloud is ok for file syncing, but not file storage. And having seen hosted providers drop things at a whim (how often have I seen, "from x date we will no longer providing y service, please download all your content prior to this date.
No thanks
like "military intelligence" or "jumbo shrimp"?
The reason I like Dropbox is that it's *fairly* reliable, with a copy in the cloud - and a copy on the hard drive in my netbook, a copy on the hard drive in my tower, and a copy in the RAM in my iPad.
But it's not nearly the size and cost to hold everything. Check the throttling on Carbonite - and the real cost - before announcing the cloud has come to destroy HDD makers.
No desktop machine can run without some sort of local storage.. whether it be SSD or spinny magnetical platters..
People don't trust cloud storage, and it's usefullness is dependant on the client's bandwidth, and the uptime of the cloud.
Gamers will NEVER cloud local installs of OS or data.. Consoles will be pushing bigger HDD's out to remove dependence on optical drives..
I'd even go as far as to suggest, that corp IT BOFHs would be very wary about putting anything other than calanders and other non-critical data on a cloud..
Clouds are over-hyped.. and, even if the HDD manufactures take a bit of a hit, I don't think it'll do them much long term damage..
If you are going to be using "the cloud" as storage, the disk makers are still going to be selling plenty of drives. To the cloud providers.
I am more or less assuming that disk usage in the cloud will even out between being able to store more than one users data on one disk, and the need for RAID and data duplication to keep everyones data.
Also i am not sure we are going to see any significant alterations in network speeds at the user level, plus with the general unreliability of the wireless networks.
"I don't keep all my cash in a hardware container in my house. Why should I keep my data in hardware containers in my house?"
Why keep your wife and kids in the house, when you can find humans in every street?
Perhaps although you are not attached to specific coins or pieces of folding paper, you are more attached to specific values of data, like those that can be rendered as photographs depicting happy memories of YOURS, the wife that married YOU etc.
Exactly. A £10 note is fungible, your sentimental photos or important documents aren't. Can you imagine being satisfied with someone saying "we can't finding your wedding photos so we've supplied you with somebody else's" or "sorry about that crucial business plan we lost but we've replaced it with another document the same size"?
This post has been deleted by its author
...it won't happen. Internet bandwidth connections are not going to be available at no less than 1Gbps (100MByte/s) for another 10 years at the very least, only a tiny percentage of the population would have far from cheap access to such bandwidth.
HDU manufacturers already have 30TB prototypes in their labs.. they are not selling those to better steal money with the whole Thailand flood fraud.
But as soon as they will start selling 10TB hard drives at $200-$300 there is no cloud service nor SSD that could match it cost, speed and reliability wise.
'Everyone everywhere has access to their 'cloud' at decent speeds ALL THE TIME'
This is naff all use to me when last week the ISP that I was using in Jordan went 'offline' for nearly half a day.
Then there is the little matter of line speeds. if my cloud is in the UK then I guess i'd get about 100kbits/sec access. This is hardly usable now is it.
I wish the people who put forward this rubbish actually got out into the real world once in a while.
Then there is the little matter of line speeds. if my cloud is in the UK then I guess i'd get about 100kbits/sec access. This is hardly usable now is it. I wish the people who put forward this rubbish actually got out into the real world once in a while.
I too get about 100kbits/second transfer rate here. This is roughly in the range of what I expect from a floppy drive. SATA 2.0 is 3 Gbit/s. It would appear likely that the writer of this article does not have a hard drive.
THE CLOUD IS BAD.
The cloud can be useful in certain limited situations despite it's handicaps.
This gets overblown into a lot of hype and nonsense as pundits ignore the limitations in current technology and gloss over the requirements of end users. It is not the idyllic Federation network from the 24th century that some people like to pretend it is.
Even if it were, one might be more inclined to store stuff with friends and family with appliances built for the purpose and made easy peasy for even the biggest n00b among us.
The very moment you place all your data in the cloud you will find that your provider doubles prices because he/she/it has you by the short and curlies.
There is NO way I will make even my personal data and applications dependent on a network connection. For a business, it's irresponsibly risky because you create a massive point of failure. If you really, really feel like doing it, install a Citrix box for 20 people and connect it via a rate restricted pipe into your network. I give it a week, tops.
(and I haven't mentioned the risk of exporting your confidential data to an untrusted 3rd party).
Intel/AMD are probably reading this and noticing that they no longer need to put expensive L1, L2, L3 cache on their processors. They can use the cloud.
A more mainstream and general purpose form of Onlive or Citrix Receiver service will probably become mainstream before the scenario outlined in the article.
> I have a Dropbox icon on my iMac
... and looking at the prices they charge, a meagre 100GB will cost you $99 per year. In antediluvian terms (i.e. before the floods in Thailand) you're buying a 1TB drive every year and only using a tenth of it. But unless you have a screamingly fast internet connection you're getting "access times" of tens of millisends - and that's presuming that (so far) few other people on your branch of the internet are contending with you for a slice of that download bandwidth. It would be even worse if you and all the other dropboxers were trying to use the same few megahertz of wifi for added wireless cool. Just wait until every bod in your road tosses they hard drives and throws money at DropBox and friends. See who'll be racing back to local hardware then!
Even if it kills consumer demand, how exactly does the author think the cloud providers are storing that data? This article strikes me as uninformed doomsaying.
Bulk data is going to go on spinning platters for the forseeable future, and whether those bits are being stored on consumer's harddrives, or cloud provider harddrives - the disk manufacturers are making money off it.
In fact, since cloud providers know their business is over if there is a high-profile data-loss incident, they're likely not using unreplicated consumer disks (while consumers, as far as I can tell, will generally use exactly that - or maybe mirrored consumer drives if they're less reckless) - so their cost per GB is higher, before even considering the expense of their data-center or their bandwidth costs. They'll of course have higher utilization than a consumer drive will, but the point is, this isn't going to destroy the disk manufacturers!
Which brings us to why it's not going to fly - consumers are cheap, and storing large amounts of data in the cloud isn't...
2TB drives can be had for $100 - half a cent per GB, and prices are back on a nice downward trend. Sales of cheap and cheerful home NAS enclosures are booming. Most of the cloud storage solutions, on the other hand, seem to be charging $0.10/GB/month. Twenty times more, in a single month, than harddrive costs, total. The annual electricity cost of running the extra HDD comes out to $5-8 per year per drive, less if you enable power-saving (4-7 watts, 8800 hrs per year -> 40-60 kWH/yr at 13 cents/kWH).
So, assuming you buy a pair of 2TB drives and run them in RAID 1 in a $100 NAS enclosure, and use half of the available capacity, you're looking at $15 of power per year and a $300 initial investment. $330 over 2 years...
Storing 1TB of data in the cloud would cost $100 per month, so over two years, the cloud user has spend a whopping $2400 - almost 8 times as much as the user with a NAS drive. Say we instead used SSDs - $350 per half tb, and put 4 of them together. $1400+200 for 4-bay NAS enclosure - it's STILL cheaper than cloud storage! And, maybe worst of all in this kind of economy, when the user with cloud storage loses his job, and he can't pay the monthly fee - there goes his data.
For the forseeable future, cloud storage is only affordable for a relatively small volume of data that needs to be synced across lots of devices (particularly mobile ones, since they can't take harddrives).
The threat to disk manufacturers isn't cloud storage, it's SSDs - they will push low capacity drives out of laptops, and push high-performance HDDs out of the enterprise, and move on from there.
I like to know where my personal stuff is. Money is not personal (especially when most of what I have isn't really mine), so I am quite happy for someone else to look after it for me. Even so, I wouldn't say no to the odd bar of gold under the floorboards...
My family pictures, personal correspondance (I store all my email locally so I can peruse it offline), financial spreadsheets and so forth, I want to know where those bits are, and know that I have control over them. It is my responsibility to look after them (backups, RAID arrays and such). I am much happier knowing that MY STUFF is in MY HOUSE and not flung far and wide in some unknown warehouse that could get flooded, go up in flames and lose everything.
Whole-heartedly agree with the general feeling of comments so far. It will be a long time indeed before my modem will compete with even my slowest spinning lump of metal, and a long time after that before it can beat an SSD.
As long as we're storing stuff in the cloud, the cloud service providers will be constantly increasing their datacentre storage ... with higher margin server grade disc drives? It's not like we'll be storing less. And almost certainly those providers will be using redundancy which most home users wouldn't have. I can't see manufacturers losing out here...
I see this as a far off outcome, with other pressures, like solid-state storage, more likely to be of more immediate concern to hard disk manufacturers. If hard drive demand plummets, I think it will be because the technology of storage itself shifts, instead of *where* the storage units live.
I see three primary barriers to a radical shift in consumers migrating storage to "the cloud". The first is convenience. If it's not fast or reliable, then average consumers will not use it as a complete replacement for local storage, no matter how carefree about the data itself. Anything that I would put in the cloud but want frequent access to needs to be small enough that I can download it in trivial amounts of time. The less often I need to access it, the bigger it can be, but there are limits. If anything that is so big it takes me hours to upload, forget it - I'm not going to fool with it. Bear in mind that most people have asymmetric upload/download speeds - I can download nearly 10x faster than I can upload.
The next barrier is trust that the provider is a long-term partner. When I put money in a bank, I don't expect that institution to shut down any time soon. Setting aside whether or not that's a safe assumption about banks in today's world, it's still one that I think most people take. I don't yet have that sense of stability with remote storage providers yet, with the possible exception of Amazon. (While some storage companies use Amazon as their underlying storage platform, that doesn't mean I would benefit fully from Amazon's corporate stability.) Also, unlike banks, storage companies aren't insured in ways that ensure our data will be returned to us or moved to another storage company should our chosen provider fail. With time, and as the industry matures and consolidates, I'm sure this will be less of a concern, but my own feeling is that is no company out there with the perceived stability that I would treat it as a permanent solution. While even a local solution has uncertainties associated with it, some can be mitigated through design, such as Drobo or other NAS products. Even if the companies that make such products go bust, I would already have their hardware on-premises, and could replace or upgrade it at a time of my choosing (barring failure), rather than be at the mercy of a distant board of directors, the economy, etc.
Finally, there is the matter of trust around the content itself, and how secure it is on a remote host. I have data on my personal systems that, if I have my way, will never see itself stored on any media that leaves my home, or at worst, perhaps a safety deposit box at a local bank. While I realize that not everyone feels this way about any or all of their data, the fact that I do means that I struggle with the idea of explicitly putting its storage in hands I cannot know. At a minimum, I would want such content encrypted end-to-end, which gets back to the point about convenience. If applying security that will make me comfortable with remote storage makes dealing with it too slow or inconvenient, then local storage is going to be more attractive.
Most users are consumers of data, and less so producers of it. Many consumers are also willing to rely on external providers to keep things they like readily available for them. This is the basis of streaming media - few copies, many views/listeners. But some of us both produce a lot of media (I toy with 3D graphics), and also dislike that external media providers may demise or otherwise remove something I like to listen to or watch. Having my own copy locally means I control my destiny, but means I must account for that storage myself, and risk loss if I have a catastrophic loss of my home. Everything has trade-offs. My preferences and choices lead me to want local storage. I suspect I'm not alone.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
