rcd@ico.isc.com (Dick Dunn) (08/02/90)
usenet@nlm.nih.gov (not a useful address) writes: > While on the subject, why do disks spin so slowly? At 3600 RPM, you > have an intrinsic delay of up to 16 msec waiting for the platter to > come around, no matter how fast the heads move. You could increase the > speed of a 5 1/4" disk by an order of magnitude without worrying about > g forces... There has been some erratic motion toward higher rotation rates, and it's become more of a trend in the last year or so. I think 5400 will be the next pseudo-standard. One of the difficulties is that if you speed up rotation, you either lose density or increase data rate. The bulk of the effort on disk drives in the last decade has gone to increasing capacity, particularly in smaller drives where there's a constraining external form factor. What's really happened is that the ability to increase linear density (useful FCPI) has grown at (*very* roughly) the same pace as the ability to handle increasing data rate...thus the rotation speed has stayed constant. Increasing data rate is a problem for two reasons: reliability in the read/write circuitry (the heads and the analog stuff out there close to them), and conformance to standards for the drive interface. There's not much sense in busting your * to get, say 4 Mb/s if you have to shove it through a 3 Mb/s hose to the CPU. Data rate increases have happened, but too slowly to allow for much change in rotation speed. In fact, you can see some high-end drives designed with dual actuators to reduce rotational latency...the tradeoff here is that it was apparently cheaper to add a second set of heads, servo, r/w logic and all, than to deal with the problems of faster rotation. -- Dick Dunn rcd@ico.isc.com -or- ico!rcd (303)449-2870 ...Software, not suits.
wayne@dsndata.uucp (Wayne Schlitt) (08/03/90)
In article <1990Aug1.220440.20727@ico.isc.com> rcd@ico.isc.com (Dick Dunn) writes: > [ ... ] In fact, you can see some high-end drives > designed with dual actuators to reduce rotational latency...the tradeoff > here is that it was apparently cheaper to add a second set of heads, servo, > r/w logic and all, than to deal with the problems of faster rotation. I have wondered about this for a long time. Are there many drives with more than one actuators? Why arent there more? I would think that this would not only reduce the rotational latency, but it could also do a lot to reduce the seek time. Depending on the smarts in the disk controller, this could make a big increase in real throughput. In the not so uncommon case of copying a file from one place on the disk to another location, you could have one actuator sitting and reading while the other one is doing the writing. Poof-da, zero seek time and possible zero rotational delay. Yeah, I know, you would have to either have two data paths to the disk or one data path that was quick real quick and a buffer so that you could keep both heads busy. You would also have to have a fast processor in order to keep the data flowing through quick enough to have zero rotational delay, but still, I could see this as a big win. -wayne
johnl@esegue.segue.boston.ma.us (John R. Levine) (08/03/90)
In article <WAYNE.90Aug2143750@dsndata.uucp> wayne@dsndata.uucp (Wayne Schlitt) writes: >I have wondered about this for a long time. Are there many drives >with more than one actuators? Why arent there more? I gather it's a mechanical problem. You typically have one surface with servo tracks that the drive uses to align the heads. If there were two sets of disk arms, you'd have to ensure that the relative positions of the corresponding heads in each set were exactly the same. As it is, the relative positions don't matter so long as the positioning is repeatable. One way to avoid that problem would be to assign the two arms to different parts of the disk, e.g. the inside half and the outside half. You could put the arms on the same actuator, which gives you the effect of half as many cylinders each with twice as many tracks, or they could have separate actuators. I believe the first case is fairly common, I've seen Fujitsu disks that do that. The second I haven't seen, probably because by the time you have two actuators, you might as well build two separate drives. -- John R. Levine, Segue Software, POB 349, Cambridge MA 02238, +1 617 864 9650 johnl@esegue.segue.boston.ma.us, {ima|lotus|spdcc}!esegue!johnl Marlon Brando and Doris Day were born on the same day.
davidsen@crdos1.crd.ge.COM (Wm E Davidsen Jr) (08/03/90)
In article <WAYNE.90Aug2143750@dsndata.uucp> wayne@dsndata.uucp (Wayne Schlitt) writes: | I have wondered about this for a long time. Are there many drives | with more than one actuators? Why arent there more? | Yeah, I know, you would have to either have two data paths to the | disk or one data path that was quick real quick and a buffer so that | you could keep both heads busy. First, you don't have to keep both heads busy. You can seek one while the other is busy, as many small systems do now with multiple disks, where there is a single data path from the controller to the main memory. The technology is pretty well in place, and seeks could be sent to the idle head while transfer was active, and to the closest head when no transfer was taking place. Second, I doubt that the electrical considerations are the problem. The form factor of disks has dropped from 14, to 8, to 5-1/4, to 3-1/2, to 2 inches (actually 5cm, the first metric disk). In those form factors you have a real problem with fitting a second actuator, and cooling it is also a problem. Some small disks specify a required aitflow as well as temperature now. You conclusions about the benefits are about right. It's not clear that the cost to benefit ratio will justify this feature at anything but the highest performance level right now. Having seen the change from the DSU10, 40 inch platters and 48MB capacity, to the 1.2GB showbox today, I think that the race for smaller may be about over, because the physical size of the drive is not much of a problem now (even at 3-1/2 inch). I predict that the next improvements will be in power and capacity, rather than just making drives tiny. -- bill davidsen (davidsen@crdos1.crd.GE.COM -or- uunet!crdgw1!crdos1!davidsen) "Stupidity, like virtue, is its own reward" -me
rcd@ico.isc.com (Dick Dunn) (08/04/90)
(I wrote) > > [ ... ] In fact, you can see some high-end drives > > designed with dual actuators to reduce rotational latency... wayne@dsndata.uucp (Wayne Schlitt) asked: > I have wondered about this for a long time. Are there many drives > with more than one actuators? Why arent there more? 1. As I pointed out, it's expensive...you need two of a lot of stuff. That rules it out for low-end ("commodity" market) drives. 2. It takes more room. Small drives are up against their form factors; it's hard to find a place to put a second actuator. 3. It gives you a nonstandard interface. If you do it, you either have to make a controller that's smart enough to take advantage of it (which, at first blush, sounds tough to me), sell enough to get people to support it, or create and sell the driver-level support yourself. 4. The mechanical aspects are tough. You have to make things solid enough that having one actuator moving doesn't drag the other one off-track. This takes some careful isolation. Watch a busy PC-class machine some time, and note how the motion of the actuator in its little drive can make the entire machine move back and forth a few millimeters. Now consider that track spacing is down in the 20-micron range...and remember that you have to be a lot closer than that to stay "on track." (You're trying to stay in your lane, where 20 microns constitutes a complete lane change.) -- Dick Dunn rcd@ico.isc.com -or- ico!rcd (303)449-2870 ...Are you making this up as you go along?
usenet@nlm.nih.gov (usenet news poster) (08/04/90)
|> rcd@ico.isc.com (Dick Dunn) writes: |> [ ... ] In fact, you can see some high-end drives |> designed with dual actuators to reduce rotational latency...the tradeoff |> here is that it was apparently cheaper to add a second set of heads, servo, |> r/w logic and all, than to deal with the problems of faster rotation. >wayne@dsndata.uucp (Wayne Schlitt) writes: >I have wondered about this for a long time. Are there many drives >with more than one actuators? Why arent there more? > >I would think that this would not only reduce the rotational latency, >but it could also do a lot to reduce the seek time. Depending on the >smarts in the disk controller, this could make a big increase in real >throughput. In the not so uncommon case of copying a file from one >place on the disk to another location, you could have one actuator >sitting and reading while the other one is doing the writing. >Poof-da, zero seek time and possible zero rotational delay. Cache busting relational databases would be another big win for this scheme. Instead of sending one head skittering all over the disk to pick up bits of the different tables as they are needed, in and ideal system you would have one head per table. Some implementations effectively do this by splitting the database across multiple disks with one table per disk, but unless you need the space, it seems wasteful. Having all the data on one disk would allow more flexibility in scheduling when the number of seek requests exceded the number of available heads. >Yeah, I know, you would have to either have two data paths to the >disk or one data path that was quick real quick and a buffer so that >you could keep both heads busy. You would also have to have a fast >processor in order to keep the data flowing through quick enough to >have zero rotational delay, but still, I could see this as a big win. If you take the "data transfer rates are the limiting factor in disk design" arguement to its logical extreme, then instead of adding more platters to increase disk capacity, you should slow the rotation rate and add more heads. Slower rotation at the same transfer rate -> higher density, more heads -> same average latency and more flexible scheduling. >-wayne David States
peter@ficc.ferranti.com (Peter da Silva) (08/04/90)
In article <2388@crdos1.crd.ge.COM> davidsen@crdos1.crd.ge.com (bill davidsen) writes: > I think that the race for smaller > may be about over, because the physical size of the drive is not much of > a problem now (even at 3-1/2 inch). I predict that the next improvements > will be in power and capacity, rather than just making drives tiny. Probably, but they won't stop making smaller drives. Not until you can put one in your wristwatch. -- Peter da Silva. `-_-' +1 713 274 5180. 'U` <peter@ficc.ferranti.com>