An Energy-Aware Combinatorial Virtual Machine Allocation and Placement Model for Green Cloud Computing


Creative Commons License

Gamsiz M., Özer A. H.

IEEE Access, cilt.9, ss.18625-18648, 2021 (SCI-Expanded) identifier identifier

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 9
  • Basım Tarihi: 2021
  • Doi Numarası: 10.1109/access.2021.3054559
  • Dergi Adı: IEEE Access
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Compendex, INSPEC, Directory of Open Access Journals
  • Sayfa Sayıları: ss.18625-18648
  • Anahtar Kelimeler: Cloud computing, Resource management, Servers, Virtual machining, Data centers, Optimization, Computational modeling, Cloud computing, combinatorial auction, energy-aware, green cloud, heuristic method, resource allocation, linear integer programming, virtual machine consolidation, virtual machine placement
  • Marmara Üniversitesi Adresli: Evet

Özet

© 2013 IEEE.Resource allocation is an important problem for cloud environments. This paper introduces an energy-aware combinatorial auction-based model for the resource allocation problem in clouds. The proposed model allows users of a cloud to submit their virtual resource requests as bids using the provided bidding language which allows complementarities and substitutabilities among those resources to be declared. The model finds the most profitable mutually satisfiable set of winning bids, and the corresponding allocation of virtual resources to the users while considering the placement of virtual resources to the available physical resources in the cloud by executing an optimization problem. During the optimization, the model also takes account of the non-linear energy requirements of the physical resources based on their utilization levels to find a placement with the lowest energy cost, thus, providing an energy-aware solution to the resource allocation problem. The associated optimization problem is formally defined and formulated using integer programming. Since the optimization problem is intractable, four heuristic methods are also proposed. To evaluate the performance of the model and the proposed heuristic methods, several experiments are conducted on a comprehensive test suite. The results demonstrate the benefits of the proposed model, and the high-quality solutions provided by the proposed methods.