Emmanuel Knafo takes a look at cloud economics and cost optimization.
In this post, we discuss Cloud Economics. We will focus on Cost Optimization which is the first pillar of the Microsoft Azure Well-Architected Framework. However, whenever possible, we will extract principles that apply to other Cloud Service Providers (CSP) such as Amazon Web Services (AWS) and Google Cloud Platform (GCP). The purpose is to cover today’s reality: many enterprises and organizations are entrenched in Multicloud and Hybrid Cloud architectures. Thus any unified and consolidated approach must encompass these cloud models.
I started my journey as Chief Information Officer (CIO) of Structube, a medium-sized furniture retailer in Canada. An all Microsoft shop with a traditional On-Premise datacenter, the natural choice for a cloud service provider is Microsoft Azure. However, like any fast growing company with relatively autonomous departments, as much as one would like to stick to a single strategic cloud provider or technology stack, inevitably, others creep in.
In any such environment, with multiple technology stacks, operating systems, and devices to manage and support, Information Technology (IT) governance is challenging to say the least. The same is true for financial aspects of IT governance (i.e. financial governance).
With the advent of cloud, terms such as commoditization have been used. While we are not interested in the polemic of whether computing has become a commodity akin to electricity, oil, gold, or wheat, we are distinctly aware that we are charged computing consumption at precise and predictable hourly rates (see Azure Pricing Calculator, Google Cloud Pricing Calculator, and AWS Pricing Calculator). Furthermore, new cloud pricing models have emerged such as reservations (see Reservations, Reserved Instances, Committed Use Discounts), spot pricing (see Azure Spot Virtual Machines, EC2 Spot Instances, Preemptible Virtual Machines), etc. In addition, CSPs offer a variety of controls, including usage quotas, budget alerts, and organizational permissions.
In this article, we will explore how to track costs from these 3 cloud service providers. We will also carve out a strategy to estimate on-premise costs. Bear in mind that on-premise cost estimations tend to be less precise (and/or accurate) and can vary greatly from one enterprise or organization to another. Tracking costs both in the cloud and on-premise in a uniform view is key: it enables us to perform cost estimation and optimization for all our workloads. This includes on-premise workloads, cloud workloads, hybrid workloads as well as workloads between clouds (i.e. intercloud workloads).
Before we dive into the details, I highly recommend listening to Adam Ronthal’s webinar “The Future of Data & Analytics is in the Cloud“. Adam Ronthal is VP Analyst at Gartner and has published many articles and webinars on financial governance for cloud among other topics. One of many key takeaways of his presentation is that the key metric you want to look at when doing cost optimization is the Price-Performance ratio.