HDD technology article

Howell

Storage? I am Storage!
Joined
Feb 24, 2003
Messages
4,740
Location
Chattanooga, TN
huh, both links go to the same article,
Hard-disk drive industry braces for technology changes
 

Groltz

My demeaning user rank is
Joined
Jan 15, 2002
Messages
1,295
Location
Pierce County, WA
To be honest, gents, the printable link takes me to the correct article only about 70% of the time.
 

jtr1962

Storage? I am Storage!
Joined
Jan 25, 2002
Messages
4,191
Location
Flushing, New York
quote from article said:
Once we master perpendicular recording, it will be the foundation for the industry up to the point of approaching 1Tbit per sq. in.

This statement is basically telling me that the HDD industry sees ~1Tbit per in² as the limit of magnetic recording. There is no mention of what technology, if any, might be suited for densities beyond that. (note:1Tbit/in² ~ 1.5 TB/platter using 3.5" platters).

Another noteworthy item is that it is getting more and more difficult to get drives with multiple platters to work. I'm guessing we've seen the last iteration of Maxtor's 4 platter platform with their 300 GB drive, and soon we'll likely see the end of 3-platter platforms as well. By the time we reach the limits of perpendicular recording I wouldn't be at all surprised if all drives are 1 platter or even 1 side. Bottom line is we'll need to go solid state to get drives of more than a few TB.
 

blakerwry

Storage? I am Storage!
Joined
Oct 12, 2002
Messages
4,203
Location
Kansas City, USA
Website
justblake.com
not sure I necesarily agree with you. I think prodicing disks with less platters is more fo a cost savings measure than anything else. I don't see why we couldn't stick with atleast 3 platter drives one the processes mature(ex: 60 is mature now, 80 will be soon).
 

jtr1962

Storage? I am Storage!
Joined
Jan 25, 2002
Messages
4,191
Location
Flushing, New York
The problem is getting the heads to switch from one disk to another for starters. The higher the areal density(and or the rotational speed) the longer this takes. Compounding this is the problem that head racks for multiple disks are just harder(i.e. more expensive) to build as areal density goes up. Sure, these problems are solved to an extent as a process matures, but there is a reason why all the drive makers but Maxtor have moved away from a four-disk platform, and just WD and Maxtor are the last hold outs with a three-disk platform. Given the areal density of today's platters, drives with more than two or three platters cater to a very niche market(and getting more niche with each increase in areal density) so it just isn't worthwhile to put much R&D into solving the attendant problems. Besides, more platters=more heat, especially as RPM increases. I'm guessing the move to 10K RPM will pretty much be the end of three-platter drives except for 7200 RPM products. Once 5400 RPM is phased out, it will doubtless mean the end of four-platter units. Further driving the industry away from multi-platter units is the fact that hard drives are now a commodity item. People are too used to spending $69 to $129 for a drive for the manufacturers to raise the prices substantially. This can be made up somewhat by increasing the number of drives sold, but in the end there's scant money left over for research. I've heard there is a greater profit margin in 3 or 4-platter units, but unless you get a certain amount of sales you won't recover the money your put into developing them.

To put things in perspective, a 2-platter drive now is 160 GB. This is already way more than 95% of users need. Maxtor's 4-platter 300 GB drive caters to a very small market. If I needed 300 GB now it would be cheaper(and probably more reliable) to just buy two smaller drives. You can buy cases with room inside for 5 or 6 drives, so unless somebody really has needs that are way out there(like 1.5 TB) I just don't see any point in purchasing anything larger than state-of-the-art two-platter drives. A two-platter platform is inherently more reliable than a three or four-platter one, and usually cheaper per GB(in fact, two-platter drives have always seemed to hit the "sweet spot" in terms of $/GB, at least since I've been buying drives). My best guess is two-platter units will be around for quite some time mainly because of the sales volume, although as we go past a few hundred GB per platter even a two-platter drive may far exceed most people's storage needs. Of course, a massive change in the way people use computers may render all this a moot point. I'm thinking that once drives become massive enough, people may have their entire video collection on their PC, and this in turn may make it worthwhile for drive manufacturers to give 3, 4, or even 5-platter drives another look. I personally won't do something like this until I can put everything on one or two drives(plus an additional drive or two to back up everything). Given that I'll need maybe 2 TB at present this looks to be a long way off. :(
 

LunarMist

I can't believe I'm a Fixture
Joined
Feb 1, 2003
Messages
16,741
Location
USA
jtr1962 said:
WD and Maxtor are the last hold outs with a three-disk platform.
You are forgetting IBM/Hitachi. ;) The three of them combined have a significant portion of the hard drive market. :) Of course their 3-platter drives most likely represent the minority of sales, since 2 platters is the sweet spot.
 

jtr1962

Storage? I am Storage!
Joined
Jan 25, 2002
Messages
4,191
Location
Flushing, New York
LunarMist said:
You are forgetting IBM/Hitachi.

They've completely slipped my mind for some reason. I usually think of Maxtor, WD, Seagate, Fujitsu, and Samsung when I think of hard drives. I guess ever since IBM's reliability fiasco they've been off my radar.
 

iGary

Learning Storage Performance
Joined
Nov 22, 2002
Messages
236
Location
iLand
jtr1962 said:
...I'm guessing we've seen the last iteration of Maxtor's 4 platter platform with their 300 GB drive, and soon we'll likely see the end of 3-platter platforms as well. By the time we reach the limits of perpendicular recording I wouldn't be at all surprised if all drives are 1 platter or even 1 side.
  • I wouldn't be at all surprised if all drives are 1 platter or even 1 side.

The sooner the better!

Once density hits a particular level, virtually all ATA (SATA) and SCSI (SAS) will transition to the 2.5-inch form factor, leaving the 3.5-inch form factor to the specialty high capacity segment (read: MaxLine III, IV and probably future competition to MaxLine from Seagate and Hitachibm).
 

Onomatopoeic

Learning Storage Performance
Joined
May 24, 2002
Messages
226
Location
LaLaLand
Ah... Well... I finally read the EE Times article.

This is the same info that the industry watchers and pundits have been pumping out semi-regularly for the past 3 or 4 years (new recording techniques).

And, yes, I did notice that they mention the upcoming transition to 2.5-inch form factor hard drives for the mainstream desktop marketplace -- so this would really be the only true news in the article.
 

Mickey

Learning Storage Performance
Joined
Mar 4, 2003
Messages
139
Location
Left Coast
blakerwry said:
not sure I necesarily agree with you. I think prodicing disks with less platters is more fo a cost savings measure than anything else. I don't see why we couldn't stick with atleast 3 platter drives one the processes mature(ex: 60 is mature now, 80 will be soon).
Going away from multiplatter platforms was partly for cost reasons and partly because of the inherent difficulty of designing for multiple configurations.

For the past few generations, one of the easiest ways of taking cost out of a drive was to remove disks and heads. With areal density increasing at close to 100% a year, it was much easier to do this than to spend time/energy designing cheaper ways of making the same parts. This is especially true when heads/media probably are the most expensive components in the drive.

Of course, one of the costs of reducing disks and heads is that media and head manufacturers started either folding or merging.

Often, to make a multiplatter drive work takes several special "tricks," what are commonly called "enablers." Maybe you have to be more sensitive to windage effects, or shock performance, or acoustics when you have more disks. These usually cost more money, either in the R&D or the component production cost, or both.

Couple the higher development costs with smaller volumes for the "large" drives, it isn't cost-effective to do so unless the profit margins are sufficient to offset the costs. There has been much debate within the industry of whether 2 or 3 disk platforms will be the long-term direction, with most firms going back and forth between the two.

As for perpendicular recording, it's been "a few years" out for the past several years. It'll probably happen someday, but predicting anything more than a year out is difficult, at best. ;)
 

blakerwry

Storage? I am Storage!
Joined
Oct 12, 2002
Messages
4,203
Location
Kansas City, USA
Website
justblake.com
mickey, is it true that on WD disks that when they get an RMA back that doesn't work, all they would re-use when making a refurbished drive would be the disk housing and the circuit board? (and not the heads/media)
 

Mickey

Learning Storage Performance
Joined
Mar 4, 2003
Messages
139
Location
Left Coast
I only know what the industry does in general, as I don't usually deal with factory operations. My understanding is when drives come back for RMA purposes, they are first scanned to confirm they really are "bad." Many are false failures which are then put into the refurb pile for sending out.

About half the industry strips out all the parts in a drive and cleans them (if they're functioning), then reuses them to build new drives. These drives are not refurb units, but prime. What always puzzled me is that since different parts can be reused different numbers of time, how do you track if you've reached that requisite cut-off? You can't mark things with a Sharpie marker, for example.

Media can't be cleaned, AFAIK, but I think there are ways to clean headstacks for reuse. Most other parts (except the motor) are pretty easy to clean, pretty much just soap and water.
 

Buck

Storage? I am Storage!
Joined
Feb 22, 2002
Messages
4,514
Location
Blurry.
Website
www.hlmcompany.com
When WD receives a drive under RMA that has already been RMAed before, it is automatically sent to Singapore for a rebuild. If this is the first time the drive has been sent in under an RMA then it is tested in various ways. If the errors are not repairable, then the product is sent to Singapore for to be rebuilt.

However, as Mickey points out, a good portion of drives that come in are NDF or No Defect Found. This is common for all hard disk drive manufacturers.
 
Top