Managing Performance Interference Effects for Intelligent and Efficient Virtual Machines Placement based on GWO Approach in Cloud

Abstract

Cloud computing paradigm has been a trend in the computational world. Thus, many service providers today are competing to enhance their features to attract more customers as they are offering them a bunch of features through a pay-as-you-go pricing model. However, despite their huge fame, cloud environments still suffer from some issues that are being studied by researchers from various perspectives. One of the controversial cloud issues nowadays is interference among virtual machines (VMs) sharing the same hardware platform called also physical machine (PM). This problem occurs due to contention on shared resources (e.g. CPU, disk, memory, network I/O, etc.) between co-hosted VMs which results in a performance degradation. The co-hosting of VMs on the same PM, emerges from the ambition of server consolidation that cloud providers aim to reach in order to improve power efficiency and optimize resource utilization. Furthermore, the Virtual Machine Placement (VMP) is one of the most challenging problems in cloud environments management and it is being studied from various perspectives. Therefore, the key factor of successful server consolidation is to minimize performance interference among co-located VMs. In this paper, we are going to review two closely related research lines (i.e. the inter-VM interference detection and/or prediction and the interference-aware virtual machine placement in cloud computing environments), give a comparative study between the reviewed approaches for each of them and propose our Swarm intelligence-based metaheuristic, named Grey Wolf Optimizer (GWO) approach, for interference aware Virtual Machine Placement Problem (VMPP).