A whale optimization system for energy-efficient container placement in data centers. (February 2021)

• Record Type: Journal Article
• Authors: Al-Moalmi, Ammar; Luo, Juan; Salah, Ahmad; Li, Kenli; Yin, Luxiu
• Is Part Of: Expert systems with applications. Volume 164(2021)
• Publication Date: 2021-02
• Subjects: Virtual machine placement -- Cloud computing -- Whale optimization -- CaaS; Expert systems (Computer science) -- Periodicals; Systèmes experts (Informatique) -- Périodiques; Electronic journals; 006.33
• DOI: 10.1016/j.eswa.2020.113719
• Languages: English
• ISSNs: 0957-4174

Abstract

Highlights: The two-level containers placement is addressed as one optimization problem. Whale optimization is adapted to address the containers placement problem. The optimization problem includes minimizing data center energy consumption. Abstract: The recent popularity of the container-as-a-service (CaaS) paradigm in data centers and with cloud providers increases the significance of the process of container deployment modeling in cloud environments. Modern data centers face the significant challenge of optimizing two objectives, power consumption and resource utilization. Thus, the task of initial placement has a new dimension, placing the containers on virtual machines (VMs) and placing these host VMs on physical machines (PMs) such that the power consumption is minimized and the resource utilization is maximized. From another perspective, the complexity of this problem increases when the heterogeneity of the containers, VMs and PMs, is considered. Therefore, in this paper, we address the problem of container and VM placement in CaaS environments with consideration of optimizing both power consumption and resource utilization. Existing solutions have addressed this problem by applying simple heuristics to the container placement problem and then applying a more sophisticated approach to the VM placement problem. In other words, the existing methods separate the two search spaces. In this work, we propose an algorithm based on the Whale Optimization Algorithm (WOA) to solve these two stages of placement as one optimization problem. The proposed algorithm searches for the optimal numbers of VMs and PMs in one search space. The proposed method is evaluated over different levels of heterogeneous environments against recent methods. Experimental results show the superiority of the proposed method over the methods of comparison on the suite of test environments.