Resource-aware parameter tuning for real-time applications

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Executing multiple applications on a multi-core system while the workload of all applications varies brings the challenge of dynamically adapting resource allocations and parametrizing with respect to constraints e.g. timing limits of real-time applications. We present a hybrid approach which extracts a set of Pareto-optimal operating points during design time which are used to dynamically parameterize the periodic application during run-time. The setup is done at the beginning of each iteration of the execution and exclusively allocates processing elements from the system depending on the current workload. The parametrization is performed with the observed information about workload complexity and allocated resources. Therefore guarantees on time limits can be granted for all iterations including situations when the number of available processing elements has been decreased sharply.

Original languageEnglish
Title of host publicationArchitecture of Computing Systems - ARCS 2019 - 32nd International Conference, Proceedings
EditorsMartin Schoeberl, Christian Hochberger, Jürgen Brehm, Thilo Pionteck, Sascha Uhrig
PublisherSpringer Verlag
Pages45-55
Number of pages11
ISBN (Print)9783030186555
DOIs
StatePublished - 2019
Event32nd International Conference on Architecture of Computing Systems, ARCS 2019 - Copenhagen, Denmark
Duration: 20 May 201923 May 2019

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11479 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference32nd International Conference on Architecture of Computing Systems, ARCS 2019
Country/TerritoryDenmark
CityCopenhagen
Period20/05/1923/05/19

Keywords

  • Parameter tuning
  • Reliability
  • Resource reservation
  • Resource-aware application
  • Self-aware application

Fingerprint

Dive into the research topics of 'Resource-aware parameter tuning for real-time applications'. Together they form a unique fingerprint.

Cite this