A Multi-Objective Ant Colony System Algorithm for Virtual Machine Placement in Traffic Intense Data Centers

• Main Authors: Yao Qin, Hua Wang, Fangjin Zhu, Linbo Zhai
• Format: Article
• Language: English
• Published: IEEE 2018-01-01
• Series: IEEE Access
• Subjects: Virtual machine placement; Ant Colony System algorithm; Pareto optimization; traffic intense data center
• Online Access: https://ieeexplore.ieee.org/document/8487034/

Virtual machine (VM) placement can meet different kinds of performance targets in data centers. As a result, it has become one of the most critical operations in data centers. In this paper, we investigate the VM placement problem for cloud applications, which have intense bandwidth requirements. In this kind of applications, all VMs communicate with a single designated point. The work of predecessors focuses on the revenue of communications in the network, and tries to find a good solution composed of the best fitted VMs. However, in their work, where to place the selected VMs is not important and has no effect on their objective, and this may cause high power consumption. We formulate the problem as a bin packing problem, which is strictly NP-hard. Then, we propose a multi-objective Ant Colony System (ACS) algorithm which is called ACS-BVMP. The goal is to obtain Pareto optimal solution set, which can simultaneously maximize the revenue of communications and minimize the power consumption of PMs. The proposed algorithm is tested with some instances from the literature. Its solution set is compared with two existing multi-objective algorithms and three single-objective algorithms. The results show that our proposed algorithm outperforms the above algorithms.