Talk:Cylinder (disk drive)

Hard disk drives are called by that name because they are not floppy (as in floppy disk drives). They are organized as a concentric stack of disks or "platters":

Each platter has two surfaces (although in practice the outer surfaces on the top and bottom of the stack are often unused because of physical space considerations), and each has its own read/write head (which reads and writes data magnetically on the surface). The data is stored on concentric circles on the surfaces known as tracks:

Corresponding tracks on all surfaces on a drive, when taken together, make up a cylinder.

Since an individual data block is one sector of a track.

By Akash Wanjari

Generality

The concept of cylinders applies to other devices than disks, e.g., data cells, drums. Also, service and timing tracks are normally not counted as parts of cylinders. Shmuel (Seymour J.) Metz Username:Chatul (talk) 18:11, 22 September 2010 (UTC)

Well, there's the physical and geometric notion of a cylinder, which I'm not sure applies to drums (a disk block is addressed in a cylindrical coordinate system, with the z coordinate selecting the track, the r coordinate selecting the cylinder, and the theta coordinate selecting the block; at least as I understand it, a drum has only a 2D coordinate system, with z and theta, but not r) or a data cell (each strip has, at least as I understand it, a 2D somewhat-Cartesian coordinate system, with one axis selecting the track and the other selecting the block - the coordinate along the first axis might be split into "cylinders" which correspond to groups of tracks and a track within that, but they're not geometrically cylinders), and there's the IBM DASD notion, which might correspond to a geometric cylinder for disks, but not, as far as I know, for drums or data cells. Guy Harris (talk) 10:19, 25 September 2010 (UTC)

On any rotating DASD with movable r/w heads there is a set of tracks that are under r/w heads for a particular position of the arm. That applies to, e.g., drums, not just to disks.

Now, for IBM DASD on the S/360 line there's the additional complication of logical cylinder. On the 2301 and 2303 drums and on the 2305 fixed-head disks, a physical cylinder would be the entire device, but for addressing and allocation purpose the devices were divided into "cylinders" that had nothing to do with geometry.

There's also the issue that I mentioned of service and support tracks that are not counted as parts of cylinders. Shmuel (Seymour J.) Metz Username:Chatul (talk) 11:14, 26 September 2010 (UTC)

Drums have arms? That's news to me....

Presumably because you're a newbie; see UNIVAC FASTRAND. I'm not aware of any other vendor with a moving head drum. Shmuel (Seymour J.) Metz Username:Chatul (talk) 11:56, 27 September 2010 (UTC)

As for fixed-head disks, if "a physical cylinder would be the entire device", that would mean that each platter has only one track. If a platter has multiple tracks, the disk has multiple cylinders, in the geometrical sense, even if that means that a cylinder can be selected purely electronically without moving an arm. Perhaps all the tracks that can be selected without moving an arm (well, modulo syncing the servo to a track - I have the impression that, in modern disks, even if you're positioned at cylinder N, if you switch tracks, the disk head might have to move in order to be centered above the track in question on the platter in question) could be called a "cylinder", in some sense, even if it's not geometrically a cylinder (e.g., in a 2321). Guy Harris (talk) 06:17, 27 September 2010 (UTC)

Cylinder (disk drive) oversimplified the definition; a cylinder is the set of all tracks accessible with a given position of the access mechanism. Usually that's a comb with one head per arm, but when you have multiple heads per arm then a cylinder is multiple concentric physical cylinders. As for track centering, that's not considered to be part of the definition of Cylinder (disk drive).Shmuel (Seymour J.) Metz Username:Chatul (talk) 11:56, 27 September 2010 (UTC)

I'd say that, at the minimum, the disk-specific notion of all tracks at a given radial position, with a one-head-per-arm access mechanism, should be mentioned, if for no other reason than to indicate why the name "cylinder" is used; for example, say that a "cylinder" for direct-access devices is the set of all tracks accessible with a given position of the access mechanism, and note that, for most disk drives, this is a cylindrical slice through the drive. Guy Harris (talk) 19:11, 27 September 2010 (UTC)

The article currently mentions the geometrical concept as motivation, and I agree that the reference should remain. I just want it to be generalized to match the actual usage of the word. I'll take a crack at it in a few days unless someone else steps up to the plate first. Shmuel (Seymour J.) Metz Username:Chatul (talk) 19:25, 27 September 2010 (UTC)