VM Power Prediction in Distributed Systems for Maximizing Renewable Energy Usage

In the context of GreenPAD project it is important to predict the energy consumption of individual (and mixture of) VMs / workload for optimal scheduling (running those VMs which require higher energy when there is more green energy available and vice-versa) in order to maximize green energy utilization. For this we execute the following experiments on an Openstack cloud testbed consisting of Fujitsu servers: VM energy measurement for different configurations (flavor + workload) and VM energy prediction for a new configuration. The automation framework for running these experiments uses bash scripts which call tools like 'stress' (simulating workloads), 'collected' (resource usage) and 'IPMI' (power measurement). We propose a linear model for predicting the power usage of the VMs based on regression. We first collect the resource usage (using collected) and the associated power usage (using IPMI) for different VM configurations and use this to build a (multi-) regression model (between resource usage and VM energy consumption). Then we use the information about the resource usage patterns of the new workload to predict the power usage. For predicting power for mix of workloads we execute (build a regression model based on) experiments with random workloads. We observe the highest energy usage for CPU-intensive workloads followed by memory-intensive workloads.