Cloud Waste is a global sustainability issue that receives little attention. It is perhaps not surprising in view of all the other global issues around climate, nature and sustainability. The total becomes too complex and overwhelming for most people to take the trouble to get to grips with it – understandably so. The more you learn, the further away the solution seems to be. So get ready for even more depressing lessons, in this blog post I will try to put Cloud Waste into this context.
First, a brief definition of Cloud Waste:
The term Cloud Waste refers to the unnecessary use of resources in the Cloud
Both individuals and companies generate Cloud Waste – in this blog post we mainly talk about waste generated by companies. There are eventually many surveys that try to estimate how much of the total investments in the Cloud is Cloud Waste. Most operate with figures from 25% to 40%. In the source references at the end of this post you will find some of these.
Economic consequences of Cloud Waste
When we talk about Cloud Waste, it is natural to compare it with a sustainability problem that has received a lot of attention in recent years; Food Waste. That issue concerns food that is produced but never eaten. UNEP, which is the UN’s program for environment and sustainability, says that the economic impact of this gigantic problem amounts to 1000 billion USD annually.
In the case of Cloud Waste, if we use a waste factor of a modest 25%, the economic consequence will amount to USD 170 billion in 2024. I have then used Gartner Group’s forecasts as the basis for the calculation. Today, the economic impact of Cloud Waste is considerably less than that of Food Waste, but that will change if we look a few years into the future. According to Gartner and IDC, the total cloud market will be 2000-2500 billion USD in 2030, with the same waste factor, the economic impact will then be 5-600 billion USD. At the same time, global commitments around Food Waste will hopefully lead to a significant improvement of this problem. It is therefore not unlikely that Cloud Waste will have economic consequences as large as Food Waste in 2030
Cloud Waste’s impact on nature and the environment
If we make a similar comparison between Cloud Waste and Food Waste in an environmental context, the outcome will not look as bad for Cloud Waste. The biggest sustainability problem related to Food Waste is the enormous areas that are seized. The official figures show that approximately 12% of the Earth’s habitable surface is seized to produce food that is never eaten. It is equivalent to the total land areas of the United States, India and Mongolia
The large, industrial data centers are significantly more space-efficient than the world’s food production. Today, they seize a total area of approximately 1,000 square kilometers, with a waste factor of 25% Cloud Waste will therefore seize 250 of these – which corresponds to an area the size of Stavanger.
However, if we look at the energy used, it looks a little worse. How much energy is used by the digital economy is a complex calculation. The most detailed review I have seen in this area was prepared by Schneider Electric in 2021. If we put this as a basis, storage, processing, data center infrastructure and networks will consume a total of 640 TWH in 2023. 25% of this is 160 TWH, which corresponds to Norway’s total electricity production.
If we are to produce 160 TWH with onshore wind turbines, there will be quite a few. A 3.5 MW onshore turbine will produce approx. 8 GWH per year. We will need 20,000 of these to produce 160 TWH of electricity. These would then occupy an area of approximately 7,000 square kilometres, i.e. considerably more than the relatively modest 250 square kilometers occupied by the data centers themselves. Still, this is not so much in the grand scheme of things, it makes up about 0.006% of the earth’s habitable surface.
When we look at the future forecasts, it immediately gets a little worse. According to the IEA, rapid adoption of energy-intensive technology will increase energy demand by between 100 TWH and 200 TWH per year, with the same waste factor of 25%, in 2030 we will waste electrical energy equivalent to 60-80,000 wind turbines on data and systems that do not create value for others than the cloud and electricity providers.
Different types of Cloud Waste
We have now established that Cloud Waste is already a significant problem, and that the future looks even worse due to the extremely rapid adoption of cloud services and explosive growth in power-intensive technology. At the same time, we see that the motivation to remove the problem should be great. The biggest consequence of Cloud Waste is financial, and the bill is paid by the consumer himself. So most companies should have plans to introduce the necessary measures, but Cloud Waste comes in several forms and can be difficult to deal with. Here we will look at the most common forms of Cloud Waste.
Poor utilization of savings plans and license benefits
This is the form of Cloud Waste that is easiest to address. Most cloud providers offer significant price discounts if the customer commits to a certain consumption over a period of time. In some cases, the savings can be over 50% compared to the so-called pay-as-you-go.
There are many reasons why these advantages are poorly utilized in a large number of companies, but it is often a matter of poor planning and an unrealistic picture of what will be achieved from digital transformation in the next year.
Inactive and overprovisioned resources
This is a significant problem for companies moving older IT systems to the cloud. Applications that are based on client/server or 3-layer architecture are in many cases moved to the cloud with the same requirements for response time and allocation of resources as in the company’s local data center. This leaves them with a continuous resource consumption that is adapted to rare high-traffic scenarios.
This problem can be solved by gaining better insight into usage patterns and capacity needs. Resources can then be allocated in line with the need for capacity, in many cases such systems can also be completely switched off for periods of time. The savings here can in many cases be over 50%
Unnecessary data and services
This is a problem that often goes under the radar. The most common variant is data that is stored for no reason. There may be old snapshots and backups of systems that have long since been cleaned up and switched off. How many backups do you need of an ERP system that is no longer in production? How long should you keep the OneDrive or home disk of users who have quit?
Other variants of this problem are user accounts that are not deleted even if the user has quit, and failed SaaS acquisitions where the agreement with the provider is not terminated even though no one uses the application.
Supplier lock-in
This is a very common problem. Most companies rely on one of the large cloud platforms for competence reasons. When the IT department, or the IT supplier, has built up expertise on one of the platforms, this is where all new services will be established. As long as the cloud providers are relatively similar in price, this will not be a major problem.
Today, most companies will have more to gain from addressing the other waste problems, but if you have everything else taken care of, you can still get 10-25% cost effectiveness by spreading different types of data and resources across the 3 big platforms ; Microsoft Azure, Amazon Web Services and Google Cloud Platform
Over-complicated architecture
A complex IT architecture can be the result of deliberate and well-reasoned actions, or a lack of control. In many large companies, the reason is both. Cost-driving architecture choices are made to create a robust and redundant infrastructure, but without the choices being anchored in stated requirements from a business that is willing to pay the bill.
Most larger companies can achieve significant cost reductions by simplifying the architecture. However, these are projects that require greater involvement from the business side.
Lack of Cloud Governance
Governance is about having the right rules and routines for the entire company, and following them up every day. Routines create culture, and a good culture creates a cost-effective business. There are three main components in a good Cloud Governance
- Requirements
What functional requirements do we set for ourselves, system and application suppliers, hired consultants and collaboration partners? - Decisions
How do we make decisions about changes and acquisitions? Who is responsible and how should we follow up the decisions? - Costs
How do we allocate and highlight cloud costs in the company?
A good Cloud Governance will have a positive impact on most forms of Cloud Waste and keep cost growth under control.
Early adopter syndrome
Finally, I want to address an issue that has the potential to become a real cost and environmental bomb. All application and system suppliers need so-called early adopters for both functional and financial reasons. Early adopters verify that the software has the right of life while ensuring the cash flow to the supplier. They therefore fulfill an important mission.
But what if a system provider could get everyone to become paying early adopters? Would they choose not to?
Software as a Service has made it easy to become an early adopter of new technology. Simply use your pre-existing Google, Facebook or Microsoft account, fill in your credit card details and suddenly you have access to shiny new technology. For several years now, we have seen a “don’t be left behind when the train leaves” polemic from large and powerful suppliers. They create a narrative that if you don’t act now, you will be outrun by competitors.
Open AI is set to reach an ARR (Annual Recurring Revenue) of USD 3.4 billion in Jun 2024. In Jun 2022, they had USD 20 million in ARR. Google and Microsoft are also creating significant growth with Gemini and Co-pilot.
It is easy to get excited by new technology from leading suppliers, and it is easy to say that we just have to have this – without making any concrete assumptions about what value the technology will create for the company, but then the result is more often Cloud Waste than value creation.
Sources used to produce this blogpost
Cloud Spend:
https://www.grandviewresearch.com/press-release/global-cloud-computing-market
https://www.fitchsolutions.com/bmi/telecommunications/what-our-clients-want-know-global-cloud-spending-outlook-2024-2030-02-04-2024
https://www.fortunebusinessinsights.com/enquiry/request-sample-pdf/cloud-computing-market-102697
https://www.digit.fyi/gartner-public-cloud-spending-to-increase-to-531-6bn-in-2024
Cloud Waste:
https://www.cloudzero.com/blog/cloud-computing-statistics
https://openmetal.io/resources/blog/control-public-cloud-waste-with-alternative-cloud
https://info.flexera.com/CM-REPORT-State-of-the-Cloud?lead_source=Organic%20Search
Cloud Environmental impact:
https://www.infoworld.com/article/3713166/questioning-clouds-environmental-impact.html
https://www.se.com/us/en/download/document/998-21202519
https://www.techrxiv.org/users/660107/articles/985442-will-energy-hungry-ai-create-a-baseload-power-demand-boom
https://www.nrk.no/ytring/et-umettelig-kraftbehov-1.16920547
https://www.wind-watch.org/faq-output.php
https://clouded.tv
Food Waste:
https://www.unep.org/news-and-stories/story/why-global-fight-tackle-food-waste-has-only-just-begun
https://www.ozharvest.org/food-waste-facts
https://www.theguardian.com/environment/ng-interactive/2015/aug/12/produced-but-never-eaten-a-visual-guide-to-food-waste
https://kitchentitbits.co.uk/2022/03/16/1-4-billion-hectares-of-land-is-used-to-produce-wasted-food
https://wwf.panda.org/discover/our_focus/food_practice/food_and_sustainable_land_use
English 
Norsk bokmål