Availability Guide for Application Design

Availability Through Process-Pairs and Monitors
Availability Guide for Application Design—525637-004
7-5
How Do Process Pairs Work?
The primary process uses interprocess communication to send critical information to
the backup process. This critical information serves two purposes:
•
The data-state information provides the backup process with enough of the primary
process’s data to continue processing without losing or corrupting data values.
•
The control-state information indicates where the backup should resume
application processing.
The backup process receives information from two sources:
•
Data-state and control-state information from the primary process
•
Messages from the operating system
Messages from the operating system indicate whether the backup needs to take over.
Takeover becomes necessary if the backup receives a message indicating that the
primary process has stopped, or that the processor the primary process is running on
has failed or is otherwise unreachable. Under these circumstances, the backup takes
over processing at the logical point indicated in the saved control-state information and
using the saved data-state information.
The next subsections expand on these generic principles to provide overviews of the
passive backup model and the active backup model of a process pair.
Figure 7-1. Process Pair—Generic Model
Primary
Control State
and Data State
$vol.appl.obj
System
messages
indicate need for
takeover
CPU 0 CPU 1
Operating
System
Operating
System
Backup
VBST601.vdd