Client: Central Government
FinOps for an essential government programme
If you are interested in FinOps and are aware of the range of services provided by Airwalk Reply, you might have already had a look at the FinOps section of this website, and you might have already read this interesting FinOps success story. Or perhaps, you have just heard about this strange new concept called FinOps, and wanted to understand a bit more.
Whatever the answer to the above, we can help you. So let’s start from the beginning...
FinOps is a discipline created by a group of cloud professionals about four years ago, with the purpose of bringing together; cloud architects, developers, finance managers, procurement and IT service owners, to increase the business value generated by the cloud operating model.
Important points about FinOps:
“I heard FinOps was about saving money”
The above statement is not entirely correct, FinOps is a practice that enables cloud professionals and businesses to manage the use of cloud to make more money.
You could have moved your workload to the cloud, adopted the FinOps principles and still be spending more than before because:
FinOps is not about reducing cost, but is about maximizing the value obtained by each unit of cost.
"We don't have enough resource to dedicate to this, how do we start our approach to FinOps?"
The FinOps Foundation suggests the following actions for a correct and complete application of FinOps principles:
It is also positive to note that every FinOps action has the additional benefit of being inline with digital sustainability objectives. This is because optimising cloud resources and workloads could potentially shrink your organisation’s carbon footprint and help meet the sustainability targets defined as part of an enterprise ESG Agenda.
Let's get back to the FinOps actions...
Imagine you are the owner of the cloud bill and the only additional information you have is an excel sheet with a list of AWS accounts, Azure subscriptions or Google Cloud projects, well, the inform action is the most effective way to provide the answer to the question of who is spending all that money?
It allows architects and engineers to identify the owner of the cloud resources and finance analysts to allocate the cloud cost to each business line, for reporting (showback), to be able to recharge it (chargeback) or to provide a detailed comparison between budgets and forecasts.
Usually, the starting point is to identify the owners of the organisational structures within the CSPs (accounts, subscriptions, projects) and assign their cost to the appropriate cost centres, and then start working on cloud tagging and labelling to obtain a higher level of granularity.
Cloud tagging and labelling are effective ways to help the FinOps team to allocate cost and classify cloud resources using metadata such as:
Cloud native cost monitoring tools (Azure Cost Monitoring, AWS Cost Explorer, Google Cloud Cost management) are recommended to start implementing the cost allocation process.
In some cases, increasing the level of resource tagging can be complicated, hence the solution for an effective allocation of the cost might be the use of in-house solutions or other third-party tools (ITSM, CCMO), available on the market. Such tools can be also very useful to solve one of the top challenges for every FinOps professional, dealing with shared cloud costs (reservations/committed use discount, networking, Kubernetes, SIEM tools, licenses, other SaaS solutions, and so on). Once the cost allocation has been consolidated, that database can be used to provide reports and different insights to every different category of stakeholder (CCoE lead, product owner, CTO, CFO).
Now that you know the cost of each project, what is the next action?
The main focus for this action is to identify opportunities to reduce the total cost of your cloud workload. Once you have started implementing your Inform processes, and you know how to use cloud native cost and workload monitoring tool, you have all you need to start optimising your cloud spend. Let me give you a couple of quick and easy examples:
You assigned the ‘Dev’ tag to a group of four Azure D4sV3 VMs and Azure monitor is showing you that they are up and running every day, and the cost monitor shows that they costs about £520 a month. Since they are development servers, and your developers are not working during the weekends, you can schedule an automated shutdown that would save you the cost of about 62 hours each week, taking the monthly bill down to £340, allowing you the budget to fund a dev server for another development team
A product team has deployed their production backend on a r7g.8xlarge EC2 machine, which provides 32 CPUs and 256GB RAM and, as shown by AWS cost explorer reports, costs approximately $1,250 per month. Analysing the server performance through AWS CloudWatch, you then realise that, in the past 30 days, the number of used CPU was constantly below six and the memory consumption never went above 64GB, hence, in agreement with the DevOps team, a rightsizing change is performed, moving to a r7g.4xlarge instance and halving the cloud bill for that single server.
Aside from the examples above there are many other optimisation patterns that can be applied, to reduce your cloud cost, for example:
Another very important pattern that could lead to cost optimisation is a periodic review and update of the existing architecture of your digital platforms, which often tend to move from more traditional, to cloud native services, improving elasticity and moving from a fixed pricing to a consumption-based pricing model.
On top of all these very reliable suggestions, you could also reduce the rates of your cloud services by buying reservations, which can give you up to 75% discount on the standard price list.
So, now that we have clarified these first two actions, we could start thinking about consolidating them into a new operating model.
The purpose of the Operate action is to define the process, the workflow, and the responsibility for carrying out the choices we made during the Optimise phase.
Now that you have an idea of how to proceed, it is time to add a bit of governance to these actions, and start defining a roadmap for your FinOps initiative. Depending on the level of maturity of your team, the stories of your FinOps epic might include:
We hope that this article gives you a better idea of what FinOps is and how to quickly start implementing some of its principles, If you need any clarification or help with some of these actions, or just want to have a chat about cloud cost optimisation, please contact us.