Browse Prior Art Database

Hardware Soft Overrun

IP.com Disclosure Number: IPCOM000046918D
Original Publication Date: 1983-Aug-01
Included in the Prior Art Database: 2005-Feb-07
Document File: 2 page(s) / 44K

Publishing Venue

IBM

Related People

Peterson, RA: AUTHOR [+2]

Abstract

During data transfer operations between a disk drive and a channel of a computer system, it is possible for an overrun to occur. An overrun occurs when the channel fails to keep up with the demand for data transfer to or from the disk drive. In the past, overruns were handled by the system under program control. That arrangement is relatively expensive and results in lower performance than the present arrangement which uses hardware to handle the overrun. When an overrun occurs, it is necessary to either back up and restart the data block in process or to hold off the starting of the data block. The disk drive is allowed to take an extra revolution to enable the channel to catch up and the operation begins at the start of the data block.

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 2

Hardware Soft Overrun

During data transfer operations between a disk drive and a channel of a computer system, it is possible for an overrun to occur.

An overrun occurs when the channel fails to keep up with the demand for data transfer to or from the disk drive. In the past, overruns were handled by the system under program control. That arrangement is relatively expensive and results in lower performance than the present arrangement which uses hardware to handle the overrun. When an overrun occurs, it is necessary to either back up and restart the data block in process or to hold off the starting of the data block.

The disk drive is allowed to take an extra revolution to enable the channel to catch up and the operation begins at the start of the data block. To facilitate the delayed start or start over on an overrun after a revolution delay, a buffer 100 (Fig. 1) for buffering at least one full data block is connected between the internal channel and the disk drive (not shown). A data block is defined as the data field on the disk drive for which an individual read or write operation can be performed. Typically called a sector, it consists of ID fields, data fields, and gap fields. The full block buffering is required because, typically, channels do not have the capability of backing up. On read operations from the disk file, it would be possible to send the first part of the data from the first read and the last part from the second read. On a write operation, without full block buffering, the first part of the data block could have been overlaid prior to the overrun and then be unav...