Dismiss
InnovationQ will be updated on Sunday, Oct. 22, from 10am ET - noon. You may experience brief service interruptions during that time.
Browse Prior Art Database

Fast Access Disk File with Several Parallel Heads

IP.com Disclosure Number: IPCOM000050480D
Original Publication Date: 1982-Nov-01
Included in the Prior Art Database: 2005-Feb-10
Document File: 3 page(s) / 103K

Publishing Venue

IBM

Related People

Blum, A: AUTHOR

Abstract

For improving the access time and data rate of a disk storage, the disks are divided into several sectors, for each of which there is one head. The bits of a byte are accessed in parallel by the various heads. After the bits have been read arbitrarily, they are electronically rearranged in the correct order by signals derived from the angular disk/head position.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 52% of the total text.

Page 1 of 3

Fast Access Disk File with Several Parallel Heads

For improving the access time and data rate of a disk storage, the disks are divided into several sectors, for each of which there is one head. The bits of a byte are accessed in parallel by the various heads. After the bits have been read arbitrarily, they are electronically rearranged in the correct order by signals derived from the angular disk/head position.

In accordance with the present proposal, the magnetic heads are associated arbitrarily, and the bits of a byte are equally distributed on a track, so that two adjacent bits on a track belong to two corresponding bits of two ascending bytes of a record. The number of permanently mounted magnetic heads on a track corresponds to the number of bits in a byte. In accordance with the nine bits of a byte (including the parity bit), nine heads are equidistantly arranged along the track, subdividing it into nine segments 0 to 8, with each segment containing one of the various bits of the stored bytes. When a read or a write operation is started, it is not necessary to wait, for example, for a match of the 0-bit string of the associated address and head 0. Instead, reading and writing are started independently of the rotational position of disk 1, so that bit 0 is accessed by head O for some operations and by another head for others. The data are simultaneously accessed by the various heads of a track, each arbitrary head accessing in parallel one arbitrary bit of a byte, so that the entire byte is accessed in a single bit access time. Thus, the data access time remaining (latency) is the disk rotation time in the sector corresponding to the associated byte address. As the different bits of a byte are accessed by different heads at different access times according to the rotational position of the disk, bit rearrangement and byte compilation are performed by electronic hardware means described later on.

Multiples of nine permanently mounted magnetic heads may be arranged on one track, thus further halving the access time and doubling the data rate, as two bytes, rather than only one, can be accessed in a single bit access time. As the very large-scale environment requires synchronously clocked high speed byte- parallel buses between the system units, two or more bytes can be transferred in parallel to and from the processor. For the purpose of parallelization, a second track can be provided and operated in parallel with the first. Because of the space requirements of heads which are permanently mounted on separate tracks, the heads can also be mechanically skewed between two or more tracks. In accordance with the geometrical and mechanical layout, the heads for one, two or more tracks can be permanently mounted on a movable arm.

Instead of equally distributing the bits of a byte on a track, it is also possible to arrange other, greater, units of information, such as bytes or blocks of bytes from a record, in the same manner.

The pres...