Member post originally published on the SparkFabrik blog
Within an ever-evolving technology landscape, the FinOps model helps to govern the rising costs of Cloud infrastructure and services. Indeed, Cloud expenses have started to rise steadily for many companies, raising the attention of those within an organisation who keep their finger on the pulse of operational costs.
If we add to this trend the increasing attention of companies to sustainability issues, both business and IT, we have the perfect situation for practices such as GreenOps to become a natural extension of companies’ operating models.
FinOps and GreenOps are clearly related methodologies, even if there is no clear-cut demarcation as to where the responsibilities of FinOps end and those of GreenOps begin. However, in pure agile thinking, it is good that the two frameworks work in concert to achieve the common goal: effective and ecological Cloud management.
FinOps: the financial management of cloud services
The FinOps Foundation defines the model this way: “FinOps is an evolving cloud financial management discipline and cultural practice that enables organisations to achieve maximum business value by helping engineering, finance, technology and business teams collaborate on data-driven spending decisions.”
The term FinOps thus refers to a discipline and cultural mindset adopted by an organisation that seeks to promote collaboration between different departments of the company in order to optimise IT spending. Specifically, the teams that are called upon by FinOps, for the benefit of the company’s cloud financial management and the business itself, are the Finance department and the DevOps teams.
This internal dialogue is geared towards maximising the value produced, balancing speed, cost and quality, without neglecting how and how much the expense (in this case of Cloud services) affects the end result. In other words, with FinOps, the end does not always justify the expense. Every decision regarding Cloud management expenditure is shared and evaluated through a data-driven decision-making process with three areas of focus*:
- Optimisation: by mapping the use of business applications
- Chargeback of costs of different applications, through the use of a tagging policy
- Cost forecasting with advanced tools that can help understand peaks and troughs in activity
(*source: PoliMi Cloud Transformation Observatory)
This approach is intended to enable teams, and the organisation itself, to progressively gain greater awareness and mastery of the optimisation levers that the Cloud provides. This leads to efficient Cloud financial management and shared financial responsibility, as we have seen in detail in the article FinOps: why move to Cloud financial management.
GreenOps: the ecological management of cloud services
Like FinOps, GreenOps is also defined as an operating model that is based on internal collaboration and leverages technologies, techniques and approaches designed to maximise efficiency in the cloud. The specific objective of GreenOps is to reduce environmental impact, measured in terms of CO2 emissions.
But does the technology sector really have a significant influence on global emissions? The answer in the affirmative comes from data from the FinOps Foundation, according to which global greenhouse gas emissions from the technology sector are comparable to or greater than those of the aviation industry, amounting to 3% and 2% of total emissions respectively. Within the 3% of emissions currently attributable to the technology sector, data centres produce about 1% of greenhouse gas emissions and consume electricity equivalent to 200-250 TWh. But the most worrying part is yet to come: according to industry analysts, if appropriate measures are not taken, emissions from the technology sector will rise to 10% of global emissions within a decade.
FinOps & GreenOps best practice
Having established that it is necessary to take measures to manage technological resources in a more sustainable manner, let us see what measures can be taken. Optimising resources and emissions can be done in various ways, some more infrastructural, others more operational. Here are nine examples of practices for green management of one’s resources:
- resource utilisation with improved cooling;
- more environmentally friendly construction materials;
- smarter control systems;
- implement the re-use of the heat and water produced;
- improve waste management;
- switch off resources during idle hours;
- choose a region that uses renewable energy (e.g. Nordic countries);
- develop an energy-efficient architecture for workloads;
- use Cloud Native solutions (e.g. serverless event-based technologies).
To learn more, we recommend you read our article on Green Cloud Computing.
FinOps and GreenOps: how to combine them for an integrated management
While FinOps allows us to keep an eye on IT spending and Cloud-related costs, GreenOps allows us to keep an eye on the grams of CO2 equivalent emitted by our infrastructure (we use the generic term infrastructure, but we talk about full Cloud or even hybrid scenarios).
The two methodologies therefore have different objectives, yet – especially in full Cloud scenarios – are closely related. A common trait between GreenOps and FinOps is the optimisation of cloud resources through proper dimensioning.
A more powerful virtual machine instance will cost more and, due to a higher power demand, will generate more emissions to run. This correlation makes it clear that it is necessary to integrate these two worlds in order to achieve the right balance between costs, emissions and required power.
Market tools such as those proposed by Cycloid and such as Kubecost, integrate the visualisation of carbon footprint data within Cost Management modules, thus giving a 360-degree view of emissions. This measurement of emissions is a process that we can define as deterministic, as it takes into account the utilisation (in a given time window) of CPU, Memory, Storage and Network and, through factors such as PUE (Power Usage Effectiveness), converts measures such as Kwh into grams of CO2 equivalent (CO2-eq). Here is a useful video to get an overview of the FinOps and GreenOps tools.
The role of Cloud Service Providers (CSPs)
It is important to understand how companies such as Google, Microsoft and Amazon intend to strive to make the operation of their data centres as sustainable and efficient as possible. In this respect, the argument becomes more complicated. We are no longer talking about emissions from underperforming or optimised software, but rather about the design of the physical infrastructure of the data centre itself. From this point of view, a Cloud Provider has three main levers at its disposal to improve its sustainability:
- PUE (Power Usage Effectiveness)
- WUE (Water Usage Effectiveness)
- Waste management
PUE (POWER USAGE EFFECTIVENESS)
PUE is a metric that measures the efficiency with which a data centre consumes and uses energy, including the operation of systems such as power, cooling, and the operation of servers, data networks and lights. The closer the PUE number is to ‘1’, the more efficient the energy use will be.
WUE (WATER USAGE EFFECTIVENESS)
Water Utilisation Effectiveness (WUE) is another key parameter and is calculated by dividing the number of litres of water used for humidification and cooling by the total annual amount of energy (measured in kWh) required to run our data centre IT equipment.
WASTE MANAGEMENT
The aim is to maximise the circularity of materials, in order to arrive as far as possible at a Zero Waste certified Data Centre (e.g. minimise single-use plastic, use dated servers for other purposes, etc.).
Benefits and challenges of integrating FinOps and GreenOps
When implementing a GreenOps and FinOps methodology, we can assume that the results will mostly be indirect or, at least, not directly perceptible to our end users. A leaner software, an optimised cloud architecture, a greater awareness of costs and how much a unit/feature costs in terms of cloud resource consumption are appreciable results, but not immediately perceptible.
The challenges, on the other hand, like any approach that requires a change of operating model and cultural change, will be much more visible to all stakeholders. We can classify them into three main areas:
- Having a clear and real visibility of emission and cost data is not always easy for organisations.
- The actual understanding of the data collected can also take time (reasoning in terms of CO2-eq is not as straightforward as comparing that same set of emissions with a number of plane trips)
- Finally, defining and implementing concrete actions to decrease costs and emissions requires effort at several levels of the organisation.
Conclusion: where to start
We have seen that both FinOps and GreenOps aim for maximum efficiency. In the first case, the focus is on avoiding the waste of financial resources; in the second, the aim is to reduce all waste in order to minimise environmental impact.
The intentions are undoubtedly useful and shareable, but there is no shortage of challenges. To reach a level of visibility and understanding of one’s infrastructure such that one has a granular and precise measurement of emissions, as well as costs, is an ambitious goal. And we can consider this as the first of the necessary steps to adopt FinOps and GreenOps methodologies. It will therefore be essential to follow a structured transformation programme, involving all departments of the organisation.