Browse Prior Art Database

Monitoring of Exceptional Conditions in a Parallel Processing System

IP.com Disclosure Number: IPCOM000113030D
Original Publication Date: 1994-Jul-01
Included in the Prior Art Database: 2005-Mar-27
Document File: 2 page(s) / 91K

Publishing Venue

IBM

Related People

Mathias, TB: AUTHOR [+4]

Abstract

Disclosed is a method of monitoring and reporting specific, definable conditions, such as error conditions and performance problems, suitable for use in a parallel processing system. The method disclosed will take advantage of the parallel processing capability to spread the workload out to the parallel processors, reducing the workload on individual processors. The goal of this method is to only report exceptional (i.e., abnormal) conditions and to avoid reporting in the absence of exceptional conditions.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 52% of the total text.

Monitoring of Exceptional Conditions in a Parallel Processing System

      Disclosed is a method of monitoring and reporting specific,
definable conditions, such as error conditions and performance
problems, suitable for use in a parallel processing system.  The
method disclosed will take advantage of the parallel processing
capability to spread the workload out to the parallel processors,
reducing the workload on individual processors.  The goal of this
method is to only report exceptional (i.e., abnormal) conditions and
to avoid reporting in the absence of exceptional conditions.

      This method consists of several main components, all of which
are interconnected through one or more communications paths.  The
first component is the monitored device (1).  These are components of
interest in a parallel processing system, such as processing units,
I/O controllers, I/O devices, etc.  Selection of "monitored devices"
will depend upon the components and events of interest.  For example,
you could consider a computer to be one device to monitor, with a
number of things to monitor, or you could consider the computer to be
a number of logical devices, each of which will be monitored.

      The second component is the monitoring device (2).  The
monitoring device will monitor one (3) or more monitored devices
(15).  It will use the interest registration table (4), which is part
of every monitoring device to determine what conditions to monitor.
To allow for a range of monitoring, there may be different classes
and subclasses (6), of items to be monitored and associated
exceptional conditions to be reported for each registered exception
processor (5).  When the monitoring device determines that a
specified condition or conditions exist in the monitored device, the
monitoring device will generate appropriate notification, based upon
the interest registration table and notify all interested exception
processor(s) (8) using the communication path(s) (9).

      In the Figu...