Seagull optimization algorithm based multi-objective VM placement in edge-cloud data centers

Abstract

Edge-Cloud Datacenters (ECDCs) have been massively exploited by the owners of technology and industrial centers to satisfy the user demand. At the same time, the amount of energy used by these data centers is considerable. To address this challenge, Virtual Machine (VM) placement of the ECDCs plays an important role; therefore, assigning VM properly to physical machines (PM) can significantly decrease the amount of energy consumption. The applied assigning technique simultaneously must consider additional objectives involving traffic and power usage of the network elements, which makes it a challenging problem. This paper proposes a multi-objective VM placement approach in edge-cloud data centers, which uses Seagull optimization to optimize power and network traffic together. In this strategy, the network traffic among PMs is reduced by concentrating the communications of VMs on the same PMs to reduce the amount of transferred data through the network and reduce the PMs’ power consumption by consolidating VMs to fewer PMs, which consumes less energy. We evaluate with simulations in CloudSim and test two different network topologies, VL2 (Virtual Layer 2) and three-tier, to validate that the proposed approach can effectively reduce traffic and power consumption in ECDCs. The experimental results show that our proposed method can decrease energy consumption by 5.5% while simultaneously reducing network traffic by 70% and the power consumption of the network components by 80%.

Keywords

• Cloud computing
• Seagull optimization
• VM placement
• Power consumption
• Network traffic
• Edge cloud data centers