Evaluating and reducing cloud waste and cost—A data-driven case study from Azure workloads

Cloud computing has been growing at an unprecedented speed in recent years. The coronavirus pandemic is further accelerating the transition of many businesses from traditional self-hosted systems to the cloud. However, an excessive number of users are struggling with over-spent budget on the cloud. Meanwhile, cloud waste is common when users provision resources beyond what they actually need. There is a clear gap in the literature to study the user behaviors in the cloud and provide viable solutions to reduce cloud cost and waste. This paper addresses these concerns by conducting a comprehensive analysis of the Microsoft Azure 2019 traces and makes the following contributions. First, we analyze the cost of 6,687 Azure cloud users, who created nearly 2.7 million virtual machines (VMs), and find that a large portion of VMs are under-utilized or over-provisioned for resources. Second, we propose Cloud Waste Points (CWP) to quantitatively evaluate the waste of each VM. We further categorize VMs that utilize cloud resource efficiently as green VMs and those that waste cloud resources as red VMs. Third, we propose Cloud Waste Indicator (CWI) as a metric to classify Azure users as red, green, and normal users, depending on their efficiency in utilizing cloud resources. In addition, we introduce Cloud Utilization Score (CUS) to rank the relative performance of Azure users in term of cloud waste. Lastly, we propose an algorithm to identify red VMs and recommend lower priced VMs that can help users reduce cost without compromising quality of service (QoS). Our experiments show that over $22 million cost savings (i.e. approximately 36% in total cost reduction) can be achieved if our recommendations are adopted by users.