Browse Prior Art Database

Statistical Technique for Detection of Exceptional Situations in MVS Systems

IP.com Disclosure Number: IPCOM000099540D
Original Publication Date: 1990-Feb-01
Included in the Prior Art Database: 2005-Mar-15
Document File: 2 page(s) / 56K

Publishing Venue

IBM

Related People

Sayer, JP: AUTHOR

Abstract

Disclosed is a statistical technique for detecting exceptional situations in MVS systems. By using this technique the time that operators must spend in determining and setting exception criteria for system monitoring can be reduced or even eliminated. The likelihood of operators spending time investigating false alarms is also reduced.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 58% of the total text.

Statistical Technique for Detection of Exceptional Situations in MVS Systems

       Disclosed is a statistical technique for detecting
exceptional situations in MVS systems.  By using this technique the
time that operators must spend in determining and setting exception
criteria for system monitoring can be reduced or even eliminated.
The likelihood of operators spending time investigating false alarms
is also reduced.

      Exceptional situations are usually detected by monitoring the
values of key system parameters and alerting the operator when any
value exceeds a predefined threshold. Statistical techniques can be
used to analyze the recent values of any system parameter and compare
each new value with its recent values.  This statistical analysis
will allow the probability of the occurrence of any particular new
value to be determined.  Exceptional situations can then be defined
as occasions when a value of a system parameter occurs that has less
than a predefined probability of occurring.

      The statistical analysis required is to determine the mean and
standard deviation of the recent values.  The difference between the
mean and any new value can be used, in conjunction with the standard
deviation, to determine the probability of the new value occurring.
The actual calculation will depend on the probability distribution of
the values of the system parameter.  An example for the case of a
system parameter that has a normal distribution is given...