In our Implementing FinOps series, we recently released the CTO’s first step i.e. understanding accrual accounting. The underlying concept of FinOps is to improve collaboration between Finance and the Office of the CTO and that is possible only with a solid conceptual understanding of the dynamics of others. In this guide, we help finance practitioners take their first step towards implementing FinOps by understanding the billing fundamentals of cloud computing. Read on!
Understanding Cloud Computing
Amazon, Microsoft, and Google are the leading providers of cloud computing services, with Amazon arguably the pioneer after having launched and expanded their Amazon Web Services (AWS) as the first-mover in the industry. But what is cloud computing?
The concept of ‘cloud computing’ is nothing more complex than this: Amazon and others have purchased thousands of computers, set them up in data centers, and developed sophisticated software that enables them to rent out access to these resources along with new, customized services created to run on them. That’s all it is. The phrase “cloud computing” was invented to describe these resources as being out “in the ether,” somewhere other than the customer’s own buildings.
Organizations benefit from cloud computing because it provides a model for enabling ubiquitous, convenient, and on-demand network access to a shared pool of configurable computing resources (e.g., networks, servers, storage, applications, and services). These resources can be rapidly provisioned and released with minimal management effort or service provider interaction. This model promotes efficiency and flexibility, scalability, and cost savings, as it allows users to increase or decrease their use of resources on an as-needed basis, paying only for what they use.
The beauty of cloud computing lies in its simplicity and the power it offers to businesses and individuals. Users can access a vast array of services and resources over the internet without needing to own or maintain the underlying infrastructure. This democratizes access to advanced computing capabilities, enabling small businesses to compete with larger ones and fostering innovation across industries.
You can read more about Cloud computing here: What is cloud computing?
Understanding the Nature and Billing Fundamentals
Although the nature of cloud computing can be summarized in fairly simple terms, the underlying technologies are incredibly complex. Amazon alone offers more than 200 different services. This number grows significantly when considering the variations within each service and adding the offerings of Microsoft and Google in multi-cloud environments. In practice, there are many thousands of cloud service SKUs available, each with its own pricing and billing structure.
The simplest services to understand from a billing standpoint would be compute and storage offerings that mimic data center infrastructure. In general, the cost to use compute instances is a function of (i) the size of the compute instance in use (larger, more capable instances being more expensive) and (ii) the period of time it is run. Similarly, the cost of storing data is normally a function of (i) how much storage is used over (ii) what time period.
Since each of these dimensions is fairly predictable, forecasting spend on these services can be relatively straightforward. However, significant complexities are introduced by newer cloud native services, which incur charges based on the variety of activities performed.
For example, the costs for using Google’s ‘Big Query‘ database depend on the amount of data scanned during a query, while costs for AWS’ ‘Redshift’ database are based on the duration of query execution and the size of the Redshift cluster configured.
The cloud providers have introduced these complexities with good intentions – namely, to cause pricing to reflect the quantity of compute resources needed to complete the customer’s activities. Accurately reflecting resource use, however, doesn’t necessarily lend itself to predictability and controllability of costs for consumers! Although cloud providers offer real-time data on costs, the final invoices at month’s end can still bring unpleasant surprises to customers’ inboxes! It is here that cloud cost management platforms come to the cloud consumer’s rescue.
How IT Procurement Changed Dramatically With Cloud!
In pre-cloud IT environments, the majority of IT procurement took place in a relatively orderly “capital expenditure” budgeting and governance model. For example, consider a scenario wherein a department within a business needs to buy 50 new servers. In most organizations, the team’s leadership would be presented with a business case justifying the outlay, and once the request is approved, a procurement team would take over the process of negotiating pricing and terms with the supplier. Once the servers were acquired and put into service, accounting for their costs would be a relatively straightforward process of accounting for depreciation over the expected life of the hardware. In pre-cloud environments, this workflow represented a comparatively straightforward process with good checks and balances.
The adoption of cloud computing services radically changes an organization’s IT procurement. Procurement of cloud services often begins with such limited planning that early adopters of cloud within the organization have been known to open their own accounts with multiple vendors and later seek reimbursement for charges incurred on their personal credit cards. In these (somewhat extreme) scenarios, the protection provided by the internal governance inherent in an orderly “business case => approval => procurement” process is lost.
Yet even when the process is more organized, for example, through the use of a central payer account across the organization and the establishment of usage guardrails, challenges remain. These measures dictate who can use cloud resources and in what quantities, but they cannot fully replicate the governance protections of the pre-cloud model. It is still the case that procurement is de facto conducted realtime all around the organization with clicks of computer mice or the execution of automated scripts. While budgets can serve as a form of governance, they cannot always prevent budget overruns from occurring.
Common Cloud Cost Management Challenges
As a result of these complications, organizations may encounter a variety of challenges in cloud cost optimization. They fall into three major categories:
Cost segmentation & allocation:
Organizations may struggle to trace costs to the individuals or teams who incurred them, and may further struggle to allocate costs to financially relevant categories such as product groupings, geographic markets, customer categories, or cost centers. These challenges can be particularly acute in the context of cloud services that are shared by multiple teams within the organization. Allocating costs incurred within containerized workloads provides a good example.
Cost overruns:
Cost overruns are often self-explanatory, yet they frequently result from a misunderstanding of how technical choices impact costs. For instance, if an organization continues to operate as in the data center, allowing non-production compute instances to run 24/7 for automatic patching, it will not only fail to take advantage of the elastic nature of cloud services, it may encounter much higher costs than before.
Other contributors to cost overruns include a failure to modernize to cloud native services or a failure to take advantage of the various discount programs offered by each cloud provider. Amazon’s Reserved Instances and Savings Plans, for example, can dramatically reduce costs on compute instances – and even further, a commitment management solution can help you save much more with minimized financial risks.
Difficulties with budgeting & forecasting:
While capital expenditure budgets in legacy IT environments were relatively predictable, forecasting cloud costs poses a challenge for technical leaders. A number of factors will cause actual outlays to vary, including the mix of resources employed, swings in realized business volumes, external environmental variations, architectural changes, and inadvertent spend incurred as a result of lapses in operator discipline. Unless a comparatively narrow and simple range of services is employed – for example basic compute instances – variances in spend relative to budgets can be a frustrating occurrence. The paradox is that use of “predictable” services such as compute instances prevents the organization from taking full advantage of the latest “cloud native” technologies. For many IT leaders this paradox can make the cloud feel like something of a Catch-22.
What is the Process of Implementing FinOps?
The good news regarding managing cloud spend is that plenty of help is available, so most of these challenges are entirely curable. Organizations, whether new to the cloud or experienced in operating cloud workloads, can establish and train a cross-functional FinOps team to address cost management challenges as they arise.
- The first step should be identifying the right mix of technical and finance professionals to serve on the team.
- As needed, these individuals can access resources on best practices from the FinOps Foundation and become FinOps certified through its training programs.
- Next, they can employ the right mix of cloud cost management tools and automated solutions to meet their needs.
An optimal combination of tooling, automation, and internal processes will allow any organization to optimize their spend while taking advantage of the elasticity and flexibility the newest cloud native technologies offer.
We hope this guide was helpful to all the finance professionals in understanding the cloud billing fundamentals. Stay tuned to our Implementing FinOps series to learn more about a smooth collaboration between Finance and Technology.
For more insights into FinOps and how to enhance your cloud cost management, visit our blog at ProsperOps or book a demo to understand how our FinOps experts can help you save more on the cloud.
About the writer: Rich Hoyer is a FinOps thought leader with over 27 years of experience in IT and Finance, and he is the CEO and co-founder of FinOptik, a niche cloud financial management agency.
