FinOps, or how to keep cloud spend under control
Reading time: 5 min
The wide use of the cloud in business can lead to a surge in costs. The FinOps approach seeks to take back control of spend by optimising resource usage in a fully efficient manner.
In an age where there is ever more infrastructure to manage, software to be deployed and updated, and data to be secured and backed up, the cloud’s promise of agility is now widely recognised. This flexibility is also reflected in the pay-per-use pricing model. However, poorly managed cloud resource usage can lead to an explosion of costs. “Projects have already been stopped or put on hold because the bill had skyrocketed,” says Hermann Dupré, global business development manager at Axians’ cloud & data centre division. “Some organisations have even backtracked and returned to on-premise infrastructure.”
“It’s about striking a balance between the agility offered by the cloud and budget constraints, and allocating the right resources in the right place, at the right time and at the right cost”
Shorthand for financial operations, FinOps provides a framework for cloud spend, ensuring that a business utilises resources in the most efficient way possible. The aim is to optimise costs wherever feasible without impacting performance or quality of service. “It’s about striking a balance between the agility offered by the cloud and budget constraints, and allocating the right resources in the right place, at the right time and at the right cost,” sums up Dupré.
Putting in place a framework and governance
FinOps is based on a framework of best practices and the creation of a governance committee bringing together IT and financial teams. This management control system for cloud spend sets out KPIs to track expenditure. It also lays down a number of rules, such as decommissioning virtual machines at the end of a project or a one-time task. For example, finance and administration departments only need computing resources at the end of the month to draw up the balance sheet, not on an on-going basis.
Another solution for controlling expenditure is to limit self-service access to IaaS (Infrastructure-as-a-Service) and PaaS (Platform-as-a-Service) by establishing a monthly rate or a validation-approval process. “With the cloud, you can create computing resources in a few clicks. This ease of use can lead to over-use by IT teams,” points out Hermann Dupré, who highlights that sensible utilisation of the cloud needs to become embedded in IT department culture. “Like digital sustainability, this issue is part of a responsible digital strategy.”
In terms of SaaS (Software-as-a-Service), FinOps aims to reduce shadow IT practices – the use by business divisions of software solutions without the approval of the IT department.
Eliminating waste and hidden costs
By centralising requests, FinOps provides an overview of services used. This means that services are paid for at the right price, and that waste and hidden costs are eliminated. While a multicloud solution can be used to select the most competitive services, provider by provider, at any given time, price structures make comparison difficult due to their deliberately complex nature.
Moreover, the cloud generates many additional costs for storage and data transfer in particular, which makes it hard to predict final spend. “At the start of a project, an organisation can also struggle to predict how much storage and bandwidth it will need,” explains Hermann Dupré.
He believes that, wherever possible, framework contracts should be implemented above a certain level of usage. A business commits for a minimum of one year to use a certain amount of resources. This fixed commitment delivers better visibility and discounted costs.
Adding AI support
Hermann Dupré considers that, in the future, part of the FinOps process should be automated with the help of artificial intelligence. By populating it with the various provider price catalogues, AI can help find resources at the best cost based on the type of request.
Take, for example, a business that needs to set up a database cluster for several days. Should it go to AWS, Microsoft Azure or Google Cloud? “There are so many factors and price structures to be taken into account that only an AI system can understand that kind of complexity,” believes the Axians expert. Time is also critical since services are charged per second or even per millisecond. An AI system could make decisions independently instead of humans.
In the meantime, the level of maturity is not the same between big businesses which have already initiated a FinOps approach and small and medium-sized companies. “There is still a need to raise awareness among these companies about these issues, as was the case for security or privacy concepts and GDPR.” Hermann Dupré notes, however, that the health crisis, by causing an explosion in the use of cloud resources, has acted as a catalyst, driving awareness.