Slashdot Mirror
Seagate Spins 15k RPM HDs
An anonymous reader sent us a story about Seagate spinning 15,000-rpm disk drive. This stuff spins faster then my head ;) I don't shop for hard drives very often... it kinda blew me away to see 40 gig IDE drives for only a few hundred bucks. I'm getting all nostalgic for the days of two 360k floppy drives. Weird.
Wow. 15 whole RPMs? Watch out, we're cruising now... That's what? About 1 Round every 4 seconds???
What Taco didn't mention, is that the platters are about four kilometers in diameter. So, 15 RPM looks pretty sweet.
I've done some reading about these new drives and word is that the super high RPMs generate so much heat that Seagate has taken to routing coolant through the drive heads just like blood through your body. Of course the little coolant pipes are tiny tiny, but they're there, chillin out, to use the parlance of our times.
Try reading the article instead of trying to get a fast post (#6 in your case). Yes, it has a faster actuator/seek time as well as faster spindle speed, 3.9 ms. The spin also brings the latency down from 2.99 ms to 2.0 ms, they said. I am curious how noisy those suckers are going to be.
And I used it in the early 1970s. Univac SS90 (90 column cards, round holes) had a 50,000 digit (5000 words x 10 digits) 17,000 rpm drum for main memory. Now Seagate is getting close, but they aren't up to that yet, and by gum, never mind that Seagate claims all the other manufacturers are 2 quarters behind: they are all 42+ years behind Univac.
Dang thing took an hour to spin down and ten minutes to spin up.
--
Infuriate left and right
Progress in storage follow Moore's Law rather closely for both size metrics (Areal density)
and performance (data rate). It is easy to over state capacity of a drive technology as the manufacture can just add more disks to up the drives capacity. The real magic is increasing the areal density of the recording media. The cost of building a drive is basically fixed. Heads cost X, disks cost Y, etc. The inductry adopts the next generation technology when it becomes cost effective.
The business model of the drive business is crazy. It take 18 months to develop a drive and it has a market life of just 6 months. The manufacturers FLY the drives to the US using cargo 747s. Also the profit margins in the drive business are razor thin.
The drive have also evolved some very cool tech over the years. If you kill power on a drive the motor becomes a generater and powers the head into the landing zone. Today's drive include either a ARM7 or 16 bit DSP class processor. As long as you don't shock the drive (1/2 ich can kill a drive) it will last forever, unlike drives of old.
Scott
The big question is how loud is this thing? I mean, the 10k drives i've experienced are pretty loud - loud when spinning and it sounds like someone knocking on the door when it's seeking... I can't imagine something even faster being any quieter.
--onyx
--onyx--
One thing that has interested me for a while is what is the chance of a HDD fragmenting. I dont mean no data defragmentation. I mean physical defragementation.
At 15,000 RPM can a HDD case contains the pieces. Even if the chance was 1 in a million I would like to know if I should put some more steel casing around my drive bays.
A friend of mine has lost feeling in his foot from a flywheel in his RX7 breaking up, it broke through the bell housing, through the steel floor then through the carpet breaking his tibia. I know it is different but a piece of HDD platter could do some serious damage.
--- Can i borrow your Clue-Stick(tm)? I need to go beat a few people with it...
Now that's fast. I remember when 60ms was fast.
Several years ago when I worked for the university, I helped throw out an old word processing system that my boss insisted was outdated (it wasn't broke, so why fix it?) It sported an old 10MB hard drive and if I remember right it was powered by a three phase motor. I laugh when a person today says installing a hard drive is complicated. Today's drive weighs less than 100 pounds and doesn't require a special circuit breaker.
Makes me want to install my advanced MFM card and see how well those state of the art IBM drives will work with my 2.2.12 kernel. Does anyone still know what RLL means anymore?
If you can't stand the risk of getting killed by your hard disk, go and play with dolls instead.
The things will be SCSI. Segate is aiming them at the Enterprise and upward class server market. Anyone who puts an IDE disk in an important server is an ignoramous. There anre a plethora of reasons to still use SCSI for server applications, although IDE is becoming very fast.
1: SCSI lets you do various EASY raid arrays.
2: The drives in a SCSI chain seek independantly of each other. The slave does not have to wait for the master.
3: More drives per buss and longer bus length.
4: Did I mention RAID already? Hot swappability rules. ten 10 GB disks beat four 25 GB's any day when you can pull a bad one out on the fly.
5: With multiple smaller disks, the data is not spread as far out on the platter, increasing seek times slightly.
I am no SCSI guru so I am sure there are more reasons!
www.mp3.com/Undocumented
If the drive heads warmed up less than the platters, the differential expansion due to thermal changes would surely distort the spacing and change the character of the way the heads ride over the platter airflow. A difference in the temperature between the air and heads could also be a source of potential problems. I have doubts they specifically cool the heads. But perhaps they do have coolant running through everything, or maybe the outer frame.
The heat sources would be the electronics (mostly underneath, but some are inside, such as the head pre-amps), the platter motor, and the voice coil. The better the bearings are, the lower the resistance to spin, and the less energy required to maintain RPM. But at higher RPM, the resistance increases by some formula I have long forgotten, so there will still me more energy needed, and thus more heat dissipated, to maintain RPM. Lighter platters would also help, but I'm not sure just to what degree this is once the drive has spun up. Head seeking needs to be faster and faster to meet our demands and expectations, too, and that means more energy in the voice coil to increase the acceleration.
So, they will be very hot! But will the heads specifically need to be cooled? I doubt it. And running coolant out to the heads would likely weight them down a whole lot.
now we need to go OSS in diesel cars
You'd think the advantages of drums would still apply today. I bet the data density and sustained read would be a lot better (and obviously more consistent) in both cases.
Are discs just that much cheaper or smaller, or what? I mean, a drum wouldn't fit nicely in the drive slot, but they might come in handy for high-performance web servers.
I bet if you had some nice solid drum drives running at that speed, you could mount them in your car and use them as flywheels for regenerative braking and to hold the world's greatest portable mp3 collection.
Platter speed is not the only concern, of course. Bit density is just as important.
There is still a long way to go. It is possible to increase the current commercial storage density by at least another order of magnitude-- I'd have to look it up. Try IBM's website, they have recently read and written in laboratory tests densities of 35.5 Gigabits per square inch.
Other technologies include ferroelectric storage (using electric polarization instead of magnetization). This has, in theory, far greater storage density than magnetic storage because the walls of ferroelectric domains are typically thousands of times thinner than those of ferromagnetic domains.
The calculation of the rotational speed that gives the maximum access speed is left as an exercise for the reader.
Not all HDDs are created equal. IBMs Deskstar 7200 RPM:ers are less noisy than most 5400 RPM:ers. And a 20+ GB drive costs about $200. Jikes! (Pun intended)
Try out fish, the friendly interactive shell.
I think it's in the jargon file; I just remember it as a bit of folklore :)
... !berkley!prime3!hawk@olivetti.atc
:)--once he saw it, he could send a message right back up the same path. If he didn't, maybe you'd be lucky and someone else would see it; maybe not.
Usenet used to rely on the arpanet backbone where available, but most sites got their feed through modems. Sending email (off arpanet) required knowing not only the destination address, but the path of every machine that the message would hop along the way (but this was easy if responding to a post; just send it back from whence it came). To email me from back east you would have sent to
something like
!lilcompanyvax!decvax1!decvax5!
gad, it's been a while; maybe I have that in the wrong direction,
and I don't remember the names of the machine, but I think that
was my final address. i
Oh, and of the 30 or so newsgroups at the time, it seems to me that two were devoted to finding paths to people. Basically, a lot of posts like, "Does anyone have a path to George Jones at Olivetti in Cupertino?" If George knew your were looking for him, he would read those newsgroups until he found your message (or grep the newsspool
Anyway, I was saying that most sites got it through modem. Then there
were the sites that didn't, which got it by tape (Australia?), leading
to the observation,
"Never underestimate the bandwith of a [station wagon|747] full of
nine track tapes."
or something like that.
/end{reminisce}