FinOps and the need for linguistic standardization.
Linguistic? That's a word and a half. Doesn't it have to do with how we speak, our language? Why couldn't I just use 'Language'?
The above is a simple example of what can go astray, especially when there are people involved, if the words we use are either not commonplace or if different people have different meanings for the same word.
Imagine this.
You hire a new FinOps Lead (let's call this Lead Jojo). One of the first things Jojo does is create cost and usage reports for every team requiring cloud resources. The reports go out, and people read them, they love the insights and are pumped about representing their teams at the upcoming FinOps Working Group session.
During the session, Infrastructure team members wholeheartedly support the desire to optimize their cloud footprint and want to know when they could begin architecting a responsive and scaleable system. This throws the Director of Finance off. "We thought optimizing cloud footprint meant reducing cost of cloud consumption, not adding to it".
This start can derail trust and confidence in any new and untried phenomenon.
Be kind to each other and sing from the same sheet.
One of the first things Jojo should have done is to start a data dictionary.
This data dictionary should have terms and jargon used frequently while discussing FinOps. Some commonly used terms are listed below.
Term | Meaning |
Cost avoidance | Either remove or rightsize a resource usage. |
Cost savings | Actual money saved because of rate and usage reduction. |
Savings potential | Potential savings that may come about as part of rate reductions or workload management. |
Savings realized | The actual amount of money saved vs what was expected. |
Rightsizing | Changing the size of provisioned resources to one that better matches need. |
Oversized/undersized | Consuming too few or too many cloud resources resulting in elevated costs or delayed operations. |
Workload management | Only consume resources when it makes sense to. Otherwise, turn off the EC2 instance or the EKS cluster. |
Commitment-based discounts | Precommitting to using a set amount of resources to get discounted rates for resource use. |
Reservation Vacancy| Unused Committment | Not using the resources that were pre-committed. You still pay for something you don't use. |
On-demand rate | Normal or base rate for using a cloud resource. |
Cost allocation (or cost attribution) | Making sure each team pays for what they used. |
I write to remember and if in the process, I can help someone learn about Containers, Orchestration (Docker Compose, Kubernetes), GitOps, DevSecOps, VR/AR, Architecture, and Data Management, that is just icing on the cake.