Affiliations: [a] The MoE Key Lab for INNS, Xi’an Jiaotong University, Xi’an, P.R. China | [b] Systems Engineering Institute, Xi’an Jiaotong University, Xi’an, P.R. China | [c] School of Computing, Electronics and Mathematics, Coventry University, Coventry, UK
| [d] Department of Genetics, Wroclaw University of Environmental and Life Sciences, Wroclaw, Poland
| [e] Department of Computer Science and Technology, Texas A&M University, Corpus Christi, TX, USA
Abstract: Consolidation of services is one of the key problems in cloud data centers. It consists of two separate but related issues: Virtual machine (VM) placement and VM migration problems. In this paper, a VM consolidation scheme is proposed that turns the virtual machine consolidation (VMC) problem into a vector packing optimization problem based on deadlock-free migration (DFM) to minimize the energy consumptions. To solve this NP-hard and computationally infeasible for large data centers problem, a novel algorithm named Chicken Swarm Optimization based on deadlock-free migration (DFM-CSO) algorithm is proposed. The DFM-CSO algorithm is characterized by the ‘one-step look-ahead with n-VMs migration in parallel (OSLA-NVMIP)’ method, which carries out the VM migration validation and the rearrangement of target physical host, as well as records the migration order for each solution placement, so that VM transfer can be completed according to the migration sequence. The experimental results, for both real and synthetic datasets, show that the proposed algorithm with higher convergence rate is favourable in comparison with the other deadlock-free migration algorithms.
Keywords: VM consolidation, VM placement, deadlock-free migration, Chicken Swarm
Optimization
