An open computing resource management framework for real-time computing

  • Authors:
  • Vuk Marojevic;Xavier Revés;Antoni Gelonch

  • Affiliations:
  • Dept. of Signal Theory and Communications, Universitat Politècnica de Catalunya, Barcelona, Spain;Dept. of Signal Theory and Communications, Universitat Politècnica de Catalunya, Barcelona, Spain;Dept. of Signal Theory and Communications, Universitat Politècnica de Catalunya, Barcelona, Spain

  • Venue:
  • HiPC'08 Proceedings of the 15th international conference on High performance computing
  • Year:
  • 2008

Quantified Score

Hi-index 0.00

Visualization

Abstract

This paper introduces an open computing resource managementframework for real-time computing systems. The framework is modular andconsists of a general computing resource modeling that facilitates a policy-based (open) computing resource management. The computing resource modelingcontains two resource model templates, which may be instantiated as oftenas necessary to capture a platform's computing resources and an application'scomputing requirements. The computing resource management approach featuresa parametric algorithm (tw-mapping with window size w) and a genericand parametric cost function, which implements the computing resource managementpolicy. We present simulations using a simple instance of this costfunction to demonstrate the suitability and versatility of the framework. Wecompute a metric that relates the computing resource management success to itscomplexity and conclude that adjusting the cost function's parameter is moreefficient than augmenting the tw-mapping's window size.