Slashdot Mirror
The Hard Business of Selling Hard Drive Platters
redfieldp writes: "This is a pretty interesting story about the 'last' HD manufacturer in the U.S., and reasons why the industry is ailing ..." There's quite a bit of interesting hard-drive history in here, too.
username: slashd0t
password: slashd0t
works
The drive market has been a commodity business for several years now. There's very little to distinguish the top offerings from the various vendors. IBM's exit from the drive arena recently was a reminder of this. A few years ago when I was part of a team designing a high-end RAID controller, it was the concensus of all the engineers that IBM made the best SCSI drives. They were dumbing billions into R&D and they still couldn't differenciate their offerings enough to make it profitable.
Here's waiting for fast solid state storage...
Jesus used to be my co-pilot, but we crashed in the mountains and I had to eat him.
The problem is they just work too well and nobody pays any attention.
:-)
I Guess we know why windows is so popular then
there has been industry speculation that Millipede is the secret advantage that led I.B.M. to decide to sell its disk-drive business to Hitachi.
I speculate it might have been due to IBM's hideous failure to manufacturing stable drives that cause them to sell out. 60% failure rate here, and thats not the floor of it!
Someone set us up the bomb, so shine we are!
Hard disk drive production capacity is far higher than demand, hence HDD manufacturers are having a harder time making a profit.
Why is this? Well three simple reasons spring to mind.
1. Current HDD capacities far exceed most users current demands.
OK, so you have more than one drive in your PC, but how many of the billion PCs sold have more than one? Servers do but they make up a very small (albeit highly profitable) segment of the HDD market. Most are installed in desktop PCs and, nowadays, most people don't use more than a fraction of the 20GB+ drives that come with a modern PC. Heck, even 5GB, the kind of capacity that was typical on an entry-level desktop three years ago is more than most users get through.
(Remember, not everyone is a MP3-fiend.)
2. We're buying fewer PCs.
Companies are buying fewer machines, as are private individuals.
Companies because the desktops that they've being buying lately need to be replaced less frequently than was previously the case (because the desktops they bought three years ago still run today's software comfortably), and because they are finding few new areas (ones that they haven't already covered) where a PC will help streamline operations. The current state of the global economy doesn't help either.
The same is essentially true for private individuals too. Anyone who wants a PC already likely has one, so why buy another one (especially in an uncertain economic climate) if the old one does the trick?
No new PC means no new HDD.
3. HDDs are now commodities.
Once something becomes ubiquitous and readily available, as HDDs have in the last five years, then it no longer demands a price premium. Fiercer competition means small profits, which means less reason to stay in the business, especially a business that ties up so much capital in the first place (in R&D and fabrication costs).
Examining these factors, especially the last one, it's not too hard to see why so many companies have exited the HDD business recently.
"Accept that some days you are the pigeon, and some days you are the statue." - David Brent, Wernham Hogg
I vote for Winchester disk.
Best Slashdot comment ever
It's just a cycle, like everything else. Hard-drives pretty much outstripped (for most people) the amount of stuff they actually store. Another thing rarely mentioned is that most people are content with what they have, not because they wouldn't like a larger hard-drive, but because it is unnecessary, and things deemed unnecessary are often the first to go when money gets tight.
On the other hand, I know of one insurance company that puts all claims and paperwork in digital form in about 4 different places. This enables them to move the paper work off site and also requires them to get the largest, most top of the line hdd's they can find. Every month or so, they are bringing a new system online with bigger, better, faster.TM So, failing harddrive companies, concentrate on the businesses, not Mom and Dad with their 12GB they wont fill up, until software bloat causes them to.
Sent from your iPad.
So how many gigs of data on your drives is actually legal?
I think that IBM's exit is about more than the marketplace being competative. I think it's about the marketplace being dead. Think about it: how much did you spend on your first 256M HD? How much does a 256M USB NVRAM "drive" cost today?
My bet is that IBM is dumping this business because it's going the way of the tape drive. Yeah, still useful for LARGE amounts of data, but it looks like it should be easy to build NVRAM drives for damn cheap, and that have a MTBF that's longer than most of us will live.
How much would it cost to build a 20G NVRAM drive that performs 10x better than a platter?
from the article
" It hasn't been easy. On Monday, Dr. Bajorek's company will announce that it is successfully emerging from Chapter 11 bankruptcy, which the company entered in May 2001."
Honestly before commenting please read the article... Companies in Chapter 11 are not traded thus they have a 0.00 dollar share price..
I refuse to argue with Anonymous Cowards - if you want a discussion get an account....
I agree that storage size has vastly outpaced demand. We have a 2 terabyte chunk of platters attached to a server which will probably triple in the next year or so, but that is not the norm.
Our "large" database servers (10's of millions of records) have more storage than they know what to with. We are currently big on 18.X gig drives at 15k rpm just beacuse we want the spindles to speed up performance. I'd rather have a 12 or 14 drive cage full of fast 18 giger ebay specials than 73 or even 36 gig drives and have a rockin price/performance ratio.
I find myself formatting drives for application servers feeling guilty that I am making partitions so big I know will never be more than a quarter full. We have web servers with less than 4 gig of space used serving about a million hits a month. Why do would we be keeping the demand up for the large drives? This drives the demand, and therefore the price and margin of the high end drives down.
The drive sizes are just growing so fast most users don't need to upgrade. It is not helped by the fact that the upgrade cycle for PC's has slowed down so much. We are replacing PC's at customers sites because the contract says it is time to replace, even though the PC is already more than powerful enough for the job they perform. How many business users really need more than a 450Mhz box on their desk? We are putting 2ghz machines on these desks now. These people run terminal emulation software, browse the web, and type.
There are many factors contributing to this hard drive problem the article talks about, these are just some personal examples I have of the reason give for the slump.
-Pete
Soccer Goal Plans
Works with this :
This indicates they don't do any manufacturing in the US? Thus are they a US manufacturer or a US owned Manufacturer ? and does this indicate there are non independant manufacturers in the US - for example IBM with US plants ? The word 'independant' is too important to be edited out of the slashdot story as it spins it in a new direction - there may be other manufacturers in the USA (i have no idea where to find out) but Komag is ONE of the last few independant ones (and i think US owned might be more valid).
This is more interesting :
So what manufacturing do they do in the US ? I suspect they have one single disk media plant and the platters are sold to OEM's for use in their drives. (they do - see Komags Website - they supply Seagate, maxtor and WD.
But in fact they don't seem to have a manufacturing plant in the us according to them - from their website
That indicates the plant that the NY Times is talking about is one of their R&D plants and not a production plant. Which it is as Komag lists San Jose and Santa Rosa as their 2 R&D plants - and for my mind R&D isn't manufacture...
So in fact are they a US manufacturer or a US owned manufacturer ? There is a difference to my mind as IBM are a US owned manufacturer.... In fact the article looks like a piece aimed at building the company's stock ahead of their relisting on the share market and not a piece about technology per se.
I refuse to argue with Anonymous Cowards - if you want a discussion get an account....
I wish this were the case, but it is generally not that American workers are over-paid, but the workers in other countries are under-paid (read exploited).
Just look at what happens...when was the last time you heard about the cheap imports from France...how about the UK...what about Germany...or Canada...or even Japan (most of the cheap electronics are made in countries like Korea and Hong Kong).
No...the truth is, it's just cheaper to buy a worker in developing/under-developed countries.
But, in the long run, it helps drive down the cost of an American worker which makes every country's workers suffer.
...Now, don't get me wrong, there are overpaid American workers, but there are also overpaid workers in most every industry and in most every developed country.
If you go to Komag's web site, you'll see that they don't make drives, they make drive platters, which they sell to drive OEMs.
General Relativity: Space-time tells matter where to go; Matter tells space-time what shape to be.
Every hard disk on the market right now has some kind of distinguishing characteristic. Folks doing equipment purchasing may not be *aware* of the distinctions, but they are present nonetheless.
Want a high-performance 5400rpm ATA disk? Look at Western Digital's *AB-series drives. Quiet SCSI? Fujitsu has/had that market cornered. Performance at any cost? Seagate's X15-36LP.
I can't say any similar thing about true commodity items like RAM or floppy disk drives. --
Visit StorageForum
-- I wanna decide who lives and who dies - Crow T. Robot, MST3K
As stated in the article, IBM had recently started to use "Pixie dust" to push the supermagnetic barrier to squeeze more data on each platter. So obviously, they ran afowl of the Pixie's union, and had to sell the business to hitachi, which relies on the gremlin's union to keep the pixies in-line.
https://www.gnu.org/philosophy/free-sw.html
From what I understand, Maxtor's engineers had reverse engineered the problem and realized that IBM was recertifying old platters stored in hungary for the newer GMR head drives. Those old platters were designed to be used in 20 GB HDs not 80 GB, so basically the problem was the same one as using a hole punch in a Single Density floppy to make it Double density, formatting it might work, but it would be far more prone to errors and data loss.
It's too bad they tarnished their reputation, but on the plus side, IBM drives are now really cheap, and a simple torture test with spinright or any program designed to contsantly overwrite the unused space on a drive should be able to punish the drive into failure, for easy replacment should it be using defective platters.
https://www.gnu.org/philosophy/free-sw.html
I wish this were the case, but it is generally not that American workers are over-paid, but the workers in other countries are under-paid (read exploited).
The economics of this sort of thing are all relative. If I lived in a third-world country and made one-tenth of my present salary, effectively I would be wealthy beyond all belief. Yeah, I'd sure be exploited!
Just look at what happens...when was the last time you heard about the cheap imports from France...how about the UK...what about Germany...or Canada
Cheap is a relative term, but costs are significantly lower for manufacturers in Canada, for instance. That is why so many American cars are made in Canada. Canadian workers are paid relatively less (or, as I said before, American workers are over-paid), and the 'artifically' low currency-exchange rate makes importing much more sensible than manufacturing in the US. You also get an educated and skilled workforce as good as or better than in the US. OTOH, the US is a big market with a big appetite for imports. Common business sense says to move cost centers (manufacturing/production) to Canada (or various other countries) and profit centers (sales) to America.
No...the truth is, it's just cheaper to buy a worker in developing/under-developed countries.
Well yeah. In a system of global economics, each country has different circumstances and can offer different comparative advantages. Third-world countries sometimes have raw natural resources to trade, but they always have cheap labour to offer.
But really, one needs to question the notion that these workers are 'exploited', given that the 'exploitation' happens on a voluntary basis. The only conclusion is that the 'exploitation' (by first-world standards) is significantly better than the alternative, presumably subsistance farming, begging, or starving to death. The anti-globalization protesters never seem to grapple with this issue. [A prosperous nation doesn't just appear overnight, and international welfare will never create one. Prosperity is the result of a long bootstrapping process that only possible under a responsible government.]
But, in the long run, it helps drive down the cost of an American worker which makes every country's workers suffer.
Bullshit. In the long run, the economics balance out to where workers are compensated in proportion to their skills and relative worth. Everybody wins in this circumstance. Perhaps American workers will ultimately stop being over-paid, but the rest of the world won't suffer because of this. With greater economic efficiency, the global standard of living increases.
but there are also overpaid workers in most every industry and in most every developed country
That's pretty much a tautology. For various political and fixed forces, some workers are paid more than they are worth, because the free market has been retarded from functioning properly.
Paraburdoo Tavern once had a sign saying `No admission without shirt and shoes. Tank-tops and thongs not acceptable' until shortly after somebody complied, turning up in a shirt and shoes. Only.
Got time? Spend some of it coding or testing
Maxtor does rock. Put simply, after that down period about six years ago, they got their sh*t together and started making a quality product, and they haven't given up. The price is right, the performance excellent (good to see Maxtor picked up Quantum's tendency to make fast-seeking IDE units.
The other really good product right now is Western Digital. They're IDE only now, unfortunately, but it take a lot of balls to stand up and recall drives from consumers, to fix a manufacturing flaw. They did it, and they earned my respect.
Samsung drives also have a really strong reputation.
In comparison we have IBM, whose last 15k SCSI unit doesn't even best Maxtor's latest 10k Atlas, and whose 7200rpm ATA models are limited by either the "Deathstar" rep or the limitations of a specificied Powered On Hours of Service specification that no one else seems to be using.
We also have Seagate, which makes some fantastic and unique products (the last 50-pin 7200rpm SCSI drive) in SCSI, and has IDE products that, frankly, suck dick. U-series drives have lousy reliability and performance that's matched by two-year old drives that are 1000rpm SLOWER. Even worse, WD's recent 5400rpm products come to wit 2% of Seagate's amazingly quite 7200rpm Barracuda IV in most benchmarks.
Most of my knowledge comes from either Storagereview.com or from Storageforum.net
-- I wanna decide who lives and who dies - Crow T. Robot, MST3K
HDs are currently the slowest thing in your computer, it is the ultimate bottleneck.
This means nothing. What about CD-Rom drives, DVD-Rom drivers, Zip drives, PCMCIA cards, Ethernet ports, USB devices, parallel ports, serial ports and floppy drives?
This is getting worse and worse each time, the performance jumps just are not present in this industry.
You are trolling big time, or you need a brain upgrade. Or perhaps simply you need to read the article. This industry's failure is that they improved way too fast. They increased the storage capacity by 100% every year! "Moore's Law? Yeah, you mean the thing we got past years ago?"
not until the manufacturers think of something creative in design
Yeah sure, those stupid morons are not creative. Every year, we tell them "There is no way you can put more data on this platter.", and every year these morons come up with new moronic ideas. Doh!
It took this long to get a 8meg cache drive, and we all know how cheap memory is.
Because of course a much bigger cache would mean a much better performance? I'm not so sure. Or else they would already have done it. You are playing a ridiculous game of "listen to me, morons". Except you're talking about very smart guys that know and take into account things you or I cannot even imagine.
There is serious lack of innovation in this field.
You seem to be a serious successful troll. Or a serious moron. You want speed? Buy several hard drives and do some RAID. You'll quickly notice that your PCI bus is very limited, though. We need 64 bits PCI cards at 66 MHz with integrated RAID controllers, and the motherboard companies are not even making them! Sheesh... There is a serious lack of innovation in the motherboard business.
Bill Gates reports that no-one will ever need more than 640KB of memory. Wait a minute - that was like two decades ago? Wow.
Black holes are where God divided by zero
Demand for storage will not increase until backup methods can scale up.
What good is that 120 GB hard drive in your machine if you can only backup 40-60 GB?
Disk storage has been really cheap for years, yet backup systems like tape and DVD are either too small or too damn expensive.
Anybody check out prices on DDS4, AIT, and other tape drives....way too expensive.
Our ability to store stuff is not dictated by hard drive space, it is dictated by backup space.
-ted
Is it just me, or does it seem there's a direct correlation between your download speeds and your hard drive size?
Unless you create huge files (digital movies) there's only so many places a person can get data to fill their hard drive. Since most DVDs are not data, the only other medium is CDs, and 650megs barely makes a dent in even a small 10gig hard drive.
So the only other source for data is the internet. If you have a 56k modem, it would take a long time to download enough to fill even 1 gig. However, with broadband, it's full in a fraction of the time. Anyone with broadband at home would agree: they downloaded a lot more once they had broadband. Whether it's mp3s, movies, games or p0rn.
Before I had broadband I had a 20 gig hard drive and couldn't even fill half of it. After broadband I bought another 20 gig, then sold them for two 40gigs. Now I'm selling those for a 120 and 100gig. All because of broadband.
If I were hard drive manufactures I'd be damn sure to market to the broadband market, either form partnerships or sell directly to customers. Because without broadband no one needs anything larger than 10gigs.