Game-Theory based Approach for Virtual Machine Placement in Cloud Data Centers
Anand Jumnal, Dilip Kumar S M · Journal of Emerging Technologies and Innovative Research · 2021
The resource on demand facility of cloud computing attracts various kinds of users to access the resources based on their Quality of Service (QoS) require- ments. The QoS differ from user to user where some users prefer time over money and vice versa. This work considers two types of users i.e. (i) deadline users (time over money) and (ii) non-deadline users (money over time) allocated to respective Virtual Machine (VM). However, managing the placements (mapping) of such different VMs into hosts (physical machines) is always an essential role in cloud data centers with objective to reduce energy usage and residual resource. This paper proposes a Game-theory approach to place two types of users’ VMs optimally in a cloud data center environment to reduce energy usage, resource wastage, and Service Level Agreement (SLA) violations. The VM migration plan considers fair migration among user VM type, SLAs, and resource utilization as well. The game is modeled with hosts as a player, VMs as a strategy, and the player’s playoff is represented by energy consumption and resource utilization. The focus of the proposed algorithm is to maximize the playoff of each player in the game to achieve the objectives. The data center’s energy consumption is minimized by the proposed algorithm, thus satisfying the Quality-of-Service (QoS) criteria of users to comply with the SLAs. The as- assessment of the proposed approach is performed using the CloudSim toolkit and compared with other participant VM placement approaches. The experimental findings indicate that the proposed approach reduces energy consumption, SLA violations while making efficient use of resources.