The direct consequence of the rapid growth of the demand for computational power by cloud based-applications has been the creation of an increasing number of large-scale data centres. In such a competitive market, each Cloud vendor needs to lower the price of the offered resources in order to increase its shares. This is done by reducing the cost associated with the execution of the users' applications, but still maintaining an adequate quality of Service. To reach this goal, each Cloud infrastructure needs to self-organise, by efficiently allocating its own resources. The complexity of the problem (exact solutions are NP-complete) calls for new, adaptive and highly-automated approaches that, at the arrival of new resource requests, are able to autonomously estimate potential resource consumptions. Hence the resource management subsystem is tuned up just keeping the associated costs as low as possible. This paper represent our contribution to this problem. We propose an approach that exploits the Cross-Entropy minimisation method to forecast the impact of different resource allocations on a Cloud infrastructure, assuming that many objective functions need to be optimised. Yet, in order to select the best allocation among those presented here, we make use of an adaptive, fast, and low resource-demanding decision-making strategy, derived from models coming from the cognitive science field. Preliminary results show the effectiveness of the proposed solution.

A Hybrid Cross-Entropy Cognitive-Based Algorithm for Resource Allocation in Cloud Environments

Dazzi P;
2014

Abstract

The direct consequence of the rapid growth of the demand for computational power by cloud based-applications has been the creation of an increasing number of large-scale data centres. In such a competitive market, each Cloud vendor needs to lower the price of the offered resources in order to increase its shares. This is done by reducing the cost associated with the execution of the users' applications, but still maintaining an adequate quality of Service. To reach this goal, each Cloud infrastructure needs to self-organise, by efficiently allocating its own resources. The complexity of the problem (exact solutions are NP-complete) calls for new, adaptive and highly-automated approaches that, at the arrival of new resource requests, are able to autonomously estimate potential resource consumptions. Hence the resource management subsystem is tuned up just keeping the associated costs as low as possible. This paper represent our contribution to this problem. We propose an approach that exploits the Cross-Entropy minimisation method to forecast the impact of different resource allocations on a Cloud infrastructure, assuming that many objective functions need to be optimised. Yet, in order to select the best allocation among those presented here, we make use of an adaptive, fast, and low resource-demanding decision-making strategy, derived from models coming from the cognitive science field. Preliminary results show the effectiveness of the proposed solution.
978-1-4799-5368-4
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/11568/1143912
 Attenzione

Attenzione! I dati visualizzati non sono stati sottoposti a validazione da parte dell'ateneo

Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 3
  • ???jsp.display-item.citation.isi??? 4
social impact