In dynamic execution contexts the achievement of desired levels of QoS requires to adapt the application conguration to statically unpredictable factors related to the physical platform and the application semantics. For distributed parallel applications this property consists in changing the parallelism degree and the parallelism form adopted by the application components. On emerging Grid and Cloud environments, recongurations often induce costs on the execution, both in terms of a performance overhead as well as a monetary charge. Therefore, advanced adaptation strategies should achieve important properties like control optimality (optimizing the application's global QoS) and recon- guration stability, expressed in terms of number of recongurations and the average time for which a conguration is not modied. In this paper we introduce a methodology based on Control Theory and Optimal Control foundations. We present a rst validation of our approach in a simulation environment, outlining its eectiveness and feasibility. Keywords: Distributed Parallel Computations, Recongurations, Autonomic Computing, Model-based Predictive Control.

Reconfiguration Stability of Adaptive Distributed Parallel Applications through a Cooperative Predictive Control Approach

MENCAGLI, GABRIELE;VANNESCHI, MARCO;VESPA, EMANUELE
2013

Abstract

In dynamic execution contexts the achievement of desired levels of QoS requires to adapt the application conguration to statically unpredictable factors related to the physical platform and the application semantics. For distributed parallel applications this property consists in changing the parallelism degree and the parallelism form adopted by the application components. On emerging Grid and Cloud environments, recongurations often induce costs on the execution, both in terms of a performance overhead as well as a monetary charge. Therefore, advanced adaptation strategies should achieve important properties like control optimality (optimizing the application's global QoS) and recon- guration stability, expressed in terms of number of recongurations and the average time for which a conguration is not modied. In this paper we introduce a methodology based on Control Theory and Optimal Control foundations. We present a rst validation of our approach in a simulation environment, outlining its eectiveness and feasibility. Keywords: Distributed Parallel Computations, Recongurations, Autonomic Computing, Model-based Predictive Control.
978-364240046-9
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: http://hdl.handle.net/11568/209055
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 6
  • ???jsp.display-item.citation.isi??? 6
social impact