Home // GREEN 2023, The Eighth International Conference on Green Communications, Computing and Technologies // View article

Cost and Carbon Reduction for Microsoft Azure Virtual Machines Using Workload Analysis

Authors:
Daisy Wong
Oliver Zhang
Jacky Huang

Keywords: cloud computing, Microsoft Azure, virtual machines, cost reduction, carbon reduction, workload analysis

Abstract:
Cloud computing has rapidly become the dominant platform for businesses across various sectors. However, many companies find it challenging to effectively control costs, often resulting in suboptimal resource allocation and unnecessary overspending. Moreover, the expansion of cloud computing has spurred a surge in electricity consumption, causing a corresponding rise in greenhouse gas emissions. This paper aims to reduce both the cost associated with Virtual Machines (VMs) in the cloud and the carbon footprint generated by cloud computing activities. To tackle this issue, we analyze the 2019 Azure cloud trace, which incorporates data from more than 2.6 million VMs and historical records of grid emissions intensity from the California ISO Northern Region. We also devise a machine learning model to predict costs based on core and memory size and formulate a waste metric that captured over 90% of the wastage in cloud workloads. In addition, we propose a cost reduction algorithm that helps to save nearly 4 million dollars. Furthermore, we developed a carbon awareness algorithm that could substantially reduce the carbon emissions of VMs by 51%.

Pages: 1 to 7

Copyright: Copyright (c) IARIA, 2023

Publication date: September 25, 2023

Published in: conference

ISSN: 2519-8483

ISBN: 978-1-68558-097-1

Location: Porto, Portugal

Dates: from September 25, 2023 to September 29, 2023