Performance of Distributed and Decentralized Control Models for Fully Distributed Processing Systems - Initial Simulation Studies.

Abstract

An essential component of a Fully Distributed Processing System (FDPS) is the distributed and decentralized control. This component unifies the management of the resources of the FDPS and provides system transparency to the user. In a previous study, the problems of distributed and decentralized control were analyzed resulting in the specification of several control models. This study continues that work by further specifying the control models defined in the first report and comparing the performance of these models in various environments. This performance analysis was accomplished by means of simulation experiments. The results of the experiments indicate that the control message traffic generated by the distributed and decentralized control is much less than expected and probably does not present a barrier to the implementation of FDPS's. Comparison of the results of the simulation of a uniprocessor and that of an FDPS indicate that little or no loss of performance is experienced by the FDPS. An important limitation of these initial performance studies is the fact that user traffic was not included in this series of tests. (Author)

Open PDF

Document Details

Document Type
Technical Report
Publication Date
May 01, 1982
Accession Number
ADA116983

Entities

People

  • Philip H. Enslow Jr.
  • Timothy G. Saponas

Organizations

  • Georgia Tech

Tags

Communities of Interest

  • Materials and Manufacturing Processes

DTIC Thesaurus Topics

  • Autonomy
  • Availability
  • Communication Systems
  • Computer Programming
  • Computer Science
  • Computers
  • Databases
  • Fault Tolerance
  • Message Processing
  • Message Systems
  • Operating Systems
  • Reliability
  • Simulations
  • Simulators
  • Specifications
  • Standards
  • Test And Evaluation

Fields of Study

  • Computer science

Readers

  • Parallel and Distributed Computing.
  • Systems Analysis and Design