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

Synchronous Multiprocessing Fail Detection

IP.com Disclosure Number: IPCOM000095853D
Original Publication Date: 1964-Aug-01
Included in the Prior Art Database: 2005-Mar-07
Document File: 2 page(s) / 42K

Publishing Venue

IBM

Related People

Dirac, JF: AUTHOR

Abstract

A pair of independently operating computer systems can be interrelated so that each can monitor the proper operation of the other. If one of the computers fails, the other computer can take over the failing computer's function.

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 96% of the total text.

Page 1 of 2

Synchronous Multiprocessing Fail Detection

A pair of independently operating computer systems can be interrelated so that each can monitor the proper operation of the other. If one of the computers fails, the other computer can take over the failing computer's function.

The two interconnected systems each comprise a central processing unit with a plurality of channels. Certain of the channels connect related control units and I/O devices. Certain channels interconnect that system to the control units of the other system. Each system has a channel connecting it to a common control unit. The latter has a storage medium for performing the monitoring of each system by the other.

At periodic intervals, each CPU registers, on the shared storage unit, certain conditions indicative of its state of operation. Between the times that one CPU registers its conditions, the other CPU registers its conditions, in an interleaved fashion. As one CPU registers its own operating state, it tests the storage medium to determine that the other CPU had registered a satisfactory state, since the last time that this other CPU had registered. In the event that the other CPU had not registered correctly, then the testing CPU scans the storage medium to determine the state of the other CPU which was apparent at the last time that the other CPU registered properly. This allows the testing CPU to take over the function of the other CPU after the point of last correct registration.

1

Page 2 of...