Disk

From Computer History Wiki
Revision as of 14:58, 22 December 2016 by Jnc (talk | contribs) (Recent developments: tyop)
Jump to: navigation, search

Disk storage was, until recently, the primary form of high-speed on-line secondary storage for computers. ('Secondary' because, unlike main memory, the data on it was not directly accessible to the CPU; rather, it has to be read into memory before the CPU can look at it.)

All disk storage is basically the same; a circular 'platter' is coated with a magnetic coating - very fine particles of magnetic material mixed with a binder, which caused the coating to adhere to the platter. A read/write head, identical in basic concept to the head on a magnetic tape recorder, is used to read and write data on the disk as the disk is rotated underneath (or above) it, by magnetizing the particles in one of two directions.

A separate mechanism is used to move the head in and out, a process known as seeking. With the head at a fixed position, it writes a circular track; the drive can write data on many tracks, each at a fixed distance from its neighbours. Many disks have more than one platter, each with its own read-write heads (usually two, one for each surface), and the collection of tracks at a given offset is referred to as a cylinder. Each track is usually divided into a number of sectors - a complete track usually contains too much data to be usefully read and written as a complete unit.

The platter rotates at high speed, and the heads are shaped to ride on a thin film of air at a very low altitude above the surface of the platter (a small spacing is required to maximize the density of the data on the platter).

Most early disk drives used packs, a group of one or more platters which could be installed or removed in a drive, in a process called mounting a pack. So a system could have multiple packs (for backup purposes, or for infrequently needed data) which could be stored until needed.

Some early disks had a separate head for each track; this avoided the delays involved in having the head(s) seek to the correct cylinder; such disks were known as 'fixed-head' disks. The platters of such disks were generally not removable.

Low-cost disks

Initially, all disks used large metal platters (usually aluminium), complex mechanisms to move the heads in and out at high speed (to reduce seek times), and were generally fairly expensive.

In the late 1960's, IBM developed a new form of disk, the floppy disk, which used instead a 'platter' of mylar (a stiff plastic); this was much cheaper, but at significantly lower performance. These disks rotate much more slowly, and the head makes contact with the surface.

The rotating substrate - not referred to as platters, but called the 'magnetic medium', or 'floppy' (although that latter term usually refers to the entire assembly) was carried in a stiff plastic envelope, with a thin felt layer on the inside (to keep the medium clean).

The first floppies were 8 inches in diameter; later they were produced in 5-1/4 inch, and finally 3-1/2 inch sizes (the latter in rigid plastic containers).

Recent developments

More recently, to increase the amount of information stored on a drive, most disk drives are now sealed. The disks in modern personal computers are generally referred to as hard drives (to distinguish them from the floppy drives which were all the earliest PC's were equipped with).