A method and system for identifying clusters of similarly-configured
computers. The method, for example, comprises gathering system and
business configuration information values from a set of computers. Next,
at least a portion of the system and business configuration information
values gathered from the set of computers is analyzed along with the
numbers of computers that are configured in various ways. Then, using the
result of the analysis as a guide, logical expressions of configuration
information values are selected for use as definitions of one or more
clusters of similarly-configured computers, the selection process
generally maximizing the number of computers included within each cluster
and also generally maximizing the number of configuration information
value specifications included within the definition of each cluster.
Then, for one or more clusters so defined and selected, one or more
statistics or limit values are generated from one or more configuration
information values or both gathered from the subset of computers included
in the one or more clusters.