Browse Prior Art Database

Managing Resource Definitions in Distributed Systems

IP.com Disclosure Number: IPCOM000116977D
Original Publication Date: 1995-Dec-01
Included in the Prior Art Database: 2005-Mar-31
Document File: 4 page(s) / 136K

Publishing Venue

IBM

Related People

Colyer, AM: AUTHOR [+6]

Abstract

A number of problems can arise when managing complex sets of resource definitions. 1. It is often desirable to keep the definitions of several resources synchronized. For example, in a parallel environment the same system configuration may need to be replicated across multiple nodes for workload balancing. Keeping these definitions in sync is a time consuming and error prone process. 2. Given a set of supposedly synchronized definitions, a user would like to view the state of those definitions in one go. Manually looking at each of the definitions in turn to verify that they are indeed the same is a time consuming and error prone process. 3. Given a definition of a single resource, a user may want to quickly replicate that definition across several systems.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 38% of the total text.

Managing Resource Definitions in Distributed Systems

      A number of problems can arise when managing complex sets of
resource definitions.
  1.  It is often desirable to keep the definitions of several
       resources synchronized.  For example, in a parallel
environment
       the same system configuration may need to be replicated across
       multiple nodes for workload balancing.  Keeping these
definitions
       in sync is a time consuming and error prone process.
  2.  Given a set of supposedly synchronized definitions, a user
would
       like to view the state of those definitions in one go.
Manually
       looking at each of the definitions in turn to verify that they
       are indeed the same is a time consuming and error prone
process.
  3.  Given a definition of a single resource, a user may want to
       quickly replicate that definition across several systems.  To
       replicate a resource definition n times a user has to log onto
n
       different systems and define that same resource n times.
  4.  Given a configuration of a system and multiple resources
defined
       on that system, a user may want to quickly replicate that
       configuration across several machines (nodes).  This currently
       requires logging onto n different nodes and defining the set
of
       resources n times.
  5.  When using a workload balancing algorithm, the user needs an
easy
       way to specify to the system the possible alternative
       destinations for a given piece of work.
  6.  In an unintelligent system where the user replicates
definitions
       with no assistance, the Enterprise Information Base (EIB) in
       which those definitions are stored can grow unnecessarily
large.
       This impacts the performance of the applications reading that
       database, and in extreme cases can cause storage problems.

      The above problems are addressed in the approach described here
by introducing the concept of Templates, Models and Clones.  These
operate within a world of configurations and definitions.
  Definition      A definition defines one real-world "thing", e.g.,
a
                  CICS*  System or a File.
  Configuration   A configuration is a set of consistent definitions.
  Model           A model is a "cookie cutter" for definitions.
                  Definitions that are created from a model are
called
                  clones.  Whenever the model is updated, the clones
                  are updated automatically.  Models are contained
                  inside configurations.
  Clone           A clone is a definition that is derived from a
model.
                  Clones are contained inside configurations.
  Template       ...