Rise of technology: not a green picture & it’s time to act
September 30, 2020
<<< Start >>>
Explore the bigger picture of Sustainable Technology and why it must be a priority on the corporate agenda by listening to our Responsible Business Days podcast with Olivier Vergeynst, founder of Green IT and Director of the Belgian Institute for Sustainable IT.
<<< End >>>
It can be argued that technology has changed our lives for the better in many ways that are also beneficial to the planet. It enables us to communicate instantly with people around the world, reducing the need to travel. It allows us to control energy usage in buildings and perform more efficient logistical planning...
However, technology comes at a price. In their study “On Global Electricity Usage of Communication Technology: Trends to 2030”, Anders Andrae and Tomas Edler estimate that IT and its entire lifecycle from build, distribution to usage is responsible for 3,8% of global greenhouse gases (GHG) emissions, well above air travel. If IT were a country, it would be the 5th biggest GHG emitter in the world. And that’s not taking into account water usage, extraction of precious minerals and e-waste.
Which leaves us where? Demand for digital technology is growing exponentially, amplified by the current COVID-19 crisis and it’s fair to presume that this upward trajectory will continue.
<<< Start >>>
As things stand today, digital transformation contributes considerably more to global warming than helping to prevent it.
<<< End >>>
Action is needed to make sustainable technology a central element in a corporate’s responsible business ambitions, with a focus on efficient use of IT resources and energy-efficient computing. Overall, research shows the biggest impact comes from end user technology equipment (PCs, mobiles, etc.). Corporates can play a role in this by looking at the different steps in the lifecycle of the equipment they own and influence, for example through their web platform or mobile application.
As a first step into the domain, companies are typically looking at a less intrusive way of making a change. Hence, in this article we take a closer look at a first critical step companies can take towards sustainable technology, optimizing their data centers.
Data centers are a critical first step for corporates to reduce their tech carbon footprint
Data is the lifeblood of digital transformation. It feeds innovation and customer experience across all sectors, especially information-intensive businesses and service industries. Data is increasing at the mindboggling rate of 2.5 quintillion bytes per day at current estimates. Along with it, the need for data centers to handle this data is rising.
<<< Start >>>
The energy used to power and cool these data centers makes them one of the top sources of GHG emissions.
<<< End >>>
They produce 200 million tons of CO2 each year and consuming 2% of electricity worldwide, projected to reach 8% by 2030 according to Anders Andrae and Tomas Edler in their study “On Global Electricity Usage of Communication Technology: Trends to 2030”. In some cases, this represents up to half of a corporate’s carbon footprint.
With more businesses relying on data-intensive computing such as artificial intelligence and big data analytics to make decisions, data centers will continue to proliferate. This makes them a critical lever to pull in terms of sustainable technology.
By adopting a Green Data Center strategy, corporates can increase energy efficiency to reduce carbon emissions and lower costs. There are three major dimensions to this: minimizing energy consumption overhead, maximizing energy utilization through tactics such as server virtualization & cloud, and optimizing energy sourcing.
1. Minimize energy consumption overhead
A large proportion of total power consumption in a data center stems from overhead, such as cooling equipment. The tool used to measure this is the Power Usage Effectiveness (PUE) metric, devised by the Green Grid, a consortium of users, technology providers, and others working to improve the resource efficiency of data centers.
<<< Start >>>
PUE is defined as the ratio between the overall power entering the data center and the power needed for the IT equipment.
<<< End >>>
A theoretically ideal PUE of 1 means that all the energy entering the data center is used to power the IT equipment and not lighting, cooling, etc.
As part of a recent client engagement, Accenture performed a PUE analysis of 13 different cases in Europe. This shows a market average PUE of 1.7, significantly higher than global cloud providers who already achieve a PUE of at least 1.2. Low PUE data centers achieve greater energy efficiency and lower energy costs. Moving to more energy efficient data centers can also deliver an estimated 26% reduction in carbon emissions.
<<< Start >>>
Natural cooling
Leading cloud providers such as Google and Microsoft are using innovative techniques to further reduce PUE. Whereas data centers are traditionally cooled using air conditioning requiring a 30 – 70% overhead in energy usage, Google’s data center facility in Hamina (Finland) is cooled with sea water. Likewise, Microsoft’s submarine data centers are naturally cooled by the sea water around them. Google is also deploying DeepMind’s AI technology to further reduce energy consumption for cooling. Using a system of neural networks trained on different operating scenarios and parameters, DeepMind has created a more efficient and adaptive framework to understand data center dynamics and optimize efficiency.
<<< End >>>
Reducing energy consumption overhead is one objective. Another is recovering overhead and putting it to good (re)use in other parts of the business. The Energy Reuse Effectiveness (ERE) metric indicates how well you reuse energy. This is defined as PUE minus energy reuse. Reusing all overhead gives an ideal ERE of 0.
<<< Start >>>
Recover & reuse
Some data centers, such as the HPC data center at the National Renewable Energy Laboratory (NREL) in Colorado (US), have gone beyond reducing overhead. They employ heat exchangers to recover the heat energy from the water used to cool the IT equipment and reuse it to heat other facilities such as offices or university campuses.
Facebook’s Danish data center plans to supply hot air waste to the district heating system of the nearby city of Odense. This will be boosted by a heat pump and delivered as hot water to citizens via a local provider. With 175,00 citizens, Odense is Denmark’s third largest city and Facebook believes it will supply up to 100,000 MWh of energy per year, enough to warm 6,900 homes.
In Stockholm (Sweden), a number of data centers are connected to the district heating system and operated by Fortum Warme, a heating company that helped to set up the Stockholm Data Parks in 2017 with the ambition to achieve low energy costs and a small environmental footprint.
<<< End >>>
2. Maximize energy utilization
Boosting the energy efficiency of data centers can also be achieved by optimizing the infrastructure itself: the servers, storage devices, and network equipment.
One way to do this is by increasing server utilization. Studies show that an idle server can consume as much as 50% of energy compared to a fully utilized server. To address this problem, virtualization technologies – using software to divide hardware-based systems into multiple isolated virtual environments – are already widely deployed in modern data centers. Server consolidation moves virtual machines among physical machines in such a way as to increase server utilization and place free physical machines in standby mode.
Dynamic Power Management (DPM) improves the power scaling abilities of servers, which reduces power drawn during idle periods. Each infrastructure resource can also be optimized on its own, e.g. the efficiency of networking equipment has been improved by reducing network port power consumption, and storage efficiency has been improved by individually increasing the storage capacity of each drive.
<<< Start >>>
Combining DPM with server virtualization and consolidation will keep physical machines busy or in a low power sleep mode in which they consume less energy and require less cooling.
<<< End >>>
Sustainability of public cloud solutions
Migrating to public cloud solutions is a sustainable choice for two reasons: it reduces individual energy consumption and the underlying resources are shared.
Cloud infrastructure leverages advanced auto-scaling mechanisms to provide computing power, storage & network capacity to companies when they request it. Virtual machines, storage and network capacity are consumed as needed rather than allocated from an on-premises set-up, the size of which is always based on peak demand.
Per definition, the underlying resources of public cloud are shared by multiple companies. This drives high utilization, allowing the reduction of a company’s own consumption to translate into a useful repurpose that cloud service providers can then sell to other companies. This in turn reduces the global data center carbon footprint.
<<< Start >>>
The efficiency & sustainability win-win offered by the cloud is a key reason for many companies to move.
<<< End >>>
A note of caution: Traditional enterprise applications are not designed to leverage the full power of the cloud. Therefore, moving to the cloud should always be combined with application modernization in order to reap the efficiency (and ecological) benefits instead of doing a ‘lift and shift’ to a cloud environment.
3. Optimize energy sourcing
A reduction in energy usage needs to go hand in hand with a strong focus on renewable energy sourcing, either procured locally, or by investing in renewables.
Again, cloud providers lead the way. For example, Amazon has a long-term commitment to achieve 100% renewable energy usage. It reached the 50% milestone in 2018, while Google has already achieved its target of 100% renewable energy.
Most cloud providers are also investing in renewable energy themselves to reach their targets. Microsoft has projects in North America, Europe and Asia to produce ~ 1.4 GW of renewable energy, in addition to leveraging locally produced green energy from wind, solar and other renewable sources. Google has installed solar farms on its data centers’ roofs, while Amazon has set up six solar farms (260 MW) and three wind farms (458 MW) in the US.
Start the journey to Sustainable Technology
Becoming a sustainable digital enterprise requires a holistic approach, where economic, social and environmental axes intertwine and inform each other. The IT function has a key role to play in this.
The journey starts with first understanding the baseline enterprise carbon footprint. Ambitions need to be formalized and priorities need to be defined based on a stakeholder & sustainability analysis. A good tool to help with this is a Sustainability Scorecard. This enables follow up and analysis of sustainability targets through measurable Objectives and Key Results (OKRs).
It is then critical to map out an action plan with concrete sustainable technology initiatives that are most relevant in the enterprise context and maximize stakeholder impact. Each initiative should have the right sponsor with clearly defined objectives on the prioritized impact levers. Data centers and their efficiency should be a crucial part of this. They are the important first step in the journey to a more sustainable digital future.
If you have any questions or would like a one-to-one chat about this topic, please contact Joeri Samyn.
