Exact and heuristic graph-coloring for energy efficient advance cloud resource reservation

Chaima Ghribi (1, 2, 3) , Djamal Zeghlache (4, 1, 2, 3)

1 TSP - RS2M - Département Réseaux et Services Multimédia Mobiles
2 SAMOVAR - Services répartis, Architectures, MOdélisation, Validation, Administration des Réseaux
3 CNRS - Centre National de la Recherche Scientifique
4 R3S-SAMOVAR - Réseaux, Systèmes, Services, Sécurité

This paper presents a new graph-coloring model for advance resource reservation with minimum energy consumption in heterogeneous IaaS cloud data centers. We start with an exact integer linear programming (ILP) formulation which generalizes the graph coloring problem and follow with a fast Energy Efficient Graph Pre-coloring (EEGP) heuristic to address the scalability and to reduce convergence times. The results of performance evaluation and comparisons of EEGP with our exact algorithm and the Haizea advance reservation (AR) algorithm demonstrate the efficiency of EEGP for the energy efficient advance resource reservation problem. Our proposed EEGP heuristic is shown to perform very close to optimal, to scale well with problem size and to achieve convergence times close to the simple and fast AR algorithm that is however suboptimal

Mots clés: Cloud computing; Heterogeneous cloud data centers; Advance resource reservation; Graph-coloring; Energy Efficiency; Computer centres; Integer programming; Linear programming; Power aware computing; Energy consumption

Domaines: Réseaux et télécommunications [cs.NI]

Chaima Ghribi, Djamal Zeghlache. Exact and heuristic graph-coloring for energy efficient advance cloud resource reservation. CLOUD 2014 : 7th International Conference on Cloud Computing, Jun 2014, Alaska, United States. pp.112 - 119, ⟨10.1109/CLOUD.2014.25⟩. ⟨hal-01281941⟩