By an anonymous FinOps Practitioner in the aviation industry
Our Public Cloud costs are now growing as we migrate from on-premise. Until now, it has been enough to use our financial planning tool to calculate and allocate shared costs across accounts tagged to business units.
Since these costs are only shown in a showback report to the business. For example, people such as Product owners do not have their own budget to use as a consumable pool of resources (we are not configured to automatically stop product resources). In effect, a single chargeback journal is done, and the backing calculations for this are very much a black box to the business users.
Our cost allocation strategy is relatively simple and reflects the way technology budgets are managed within our organization by finance teams.
The total infrastructure cost is viewed as a single budget line item entry. This cost is charged back by a central IT Finance function to each business unit that consumes their applications, resources, and shared infrastructure. These are charged based on agreed allocation rules which are different from on-premise infrastructure (CAPEX).
With the migration to the public cloud, these rules, processes, and cost allocations have now moved to OPEX. Individual applications are allocated by isolating the space in which the application is processed. However, this will change when tagging policies and processes are implemented, which will also support security rules and compliance.
A robust tagging framework will enable more detailed reporting of service consumption and allocation rules for cost showback and, finally, budget assignment.
While our Public Cloud costs are low, these processes are acceptable and status quo. As the costs increase, there will be a need for end-to-end showback and chargeback methods for product owners to have visibility.
Looking ahead, we hope to manage budgets at a more granular level to reduce the impact of central rate commitments by allocating the costs of commitments for those services that benefit based on the cloud provider’s algorithms. We are currently thinking about maintaining a simple approach to all of this while ensuring we incentivize the right behaviors from teams.
