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

Rapid Reactivation of a Backup Processor

IP.com Disclosure Number: IPCOM000035688D
Original Publication Date: 1989-Aug-01
Included in the Prior Art Database: 2005-Jan-28
Document File: 3 page(s) / 105K

Publishing Venue

IBM

Related People

Griscom, RE: AUTHOR [+2]

Abstract

In a two-way redundant processor arrangement, the processors are typically configured such that one is active and the other is backup. When the active processor fails, the backup takes over by first reinitializing itself as active. The reinitialization time for the backup to become active constitutes a disruption in the net processing of the pair of processors because no output is produced during that time. This article describes two types of improvements to this design, each of which reduces the time of disruption. (Image Omitted)

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 57% of the total text.

Page 1 of 3

Rapid Reactivation of a Backup Processor

In a two-way redundant processor arrangement, the processors are typically configured such that one is active and the other is backup. When the active processor fails, the backup takes over by first reinitializing itself as active. The reinitialization time for the backup to become active constitutes a disruption in the net processing of the pair of processors because no output is produced during that time. This article describes two types of improvements to this design, each of which reduces the time of disruption.

(Image Omitted)

In Design "A" of Fig. 1, the program initialization sequence of the processors is partitioned into two sections, "init1" and "init2". "Init1" performs all initialization which is independent of the inputs to the system or of the state of the active processor. "Init1" is performed on each processor, whether it is going to be active or backup. On the backup, this has the effect of priming the backup for takeover. Since "Init1" is performed prior to the time that the backup is needed as an active, it therefore does not contribute to the net system disruption when the backup does become active.

"Init2" performs only initialization which is dependent upon the inputs and the state of the active processor. "Init2" is only done on a processor that is going to be active. When executed for recovery purposes, 'init2's actions may be based on a recovery vector provided in hardware, such as footprints. In pr...