Digitalization is a game changer in the pursuit of a sustainable future. The latest digital technologies and applications like cloud, AI, and mobile devices enable us to achieve the Sustainable Development Goals and reduce carbon emissions in many sectors. Yet computer systems themselves have an immense energy requirement for their countless devices, data centers, applications and global networks. To effectively reduce the carbon footprint of digitalization, it is necessary to apply algorithmic efficiency and sustainability by design as guiding principles in digital engineering.
The clean-IT Forum is the international platform to exchange ideas, recent research findings and applications to make digital technologies more energy-efficient.
This course differs from the common openHPI format. There are no deadlines or graded tests. The clean-IT Forum spotlights the pressing issue of the impact of digitization on climate change and how to deliver sustainable digital technologies. Experts from various organizations will share their ideas about how to make digital technologies more sustainable and will be open to discussing with you how to spread clean-IT solutions. You are not simply a learner, but may contribute your own ideas about how to reduce the energy consumption of computer systems by sharing your videos on the platform and your ideas in the discussion forum.
The clean-IT Forum is therefore an open platform for stakeholders in academia, IT industry, NGOs and policy leaders to exchange ideas about how to effectively reduce the energy requirements of digital technologies following the guiding principles "algorithmic efficiency" and "sustainability by design" in digital engineering. Due to their immateriality computer systems and software appear to be clean services and applications. However, every digital action requires energy. The use of digital technologies has increased exponentially over the last couple of decades and therefore also its energy requirements - this is especially true for such popular applications as cloud, artificial intelligence and media streaming.
To reduce the energy requirements of computer systems it is necessary to:
- Raise awareness about the energy footprint of computer systems
- Find feasible methods to measure the energy consumption of computer systems and software
- Take the trade-off between performance and energy consumption into account when creating computer systems
- Establish algorithmic efficiency and sustainability by design as guiding principles in digital engineering
- Rethink IT architectures and algorithms
- Apply clean-IT solutions on a broad scale in popular services and products
The clean-IT Forum will shed light on all of the aforementioned steps and present clean-IT solutions from research and application in very different domains of digital technologies by supporters of the clean-IT initiative:
. Digitalization and Climate
. Energy-efficient Algorithms and Databases
. Energy-efficient Data Centers
. Clean AI
The clean-IT Forum is open to everyone. If you want to share your clean-IT insight or solution, let us know and get started!
Course contents
Digitalization and Climate:
Digital technologies help to reduce the carbon emissions massively on a global scale. They are indispensable to achieving the Sustainable Development Goals. However, digital systems have their own share of carbon emissions, which are steadily growing. How much do computer systems add to the global carbon footprint? How much energy do digital systems require and what measures have already been taken to try to reduce this? What can we do in everyday life to reduce energy consumption of digital devices and services? What is missing so far and how can clean-IT and sustainability by design help to further decrease the carbon footprint of computer systems?
Energy-efficient Algorithms and Databases:
All digital technologies are based on algorithms and data. Metaphorically speaking, computer systems are like alchemy. Very few things are impossible in the digital world. Any scenario or program can be modelled by sequences of binary values (0-1). However, the quality of digital programs varies significantly. One can achieve the same result with very different algorithmic sequences. A large, complex and wastefully-designed piece of software can facilitate the same outcome as a slim, simple and thoughtfully-designed algorithm. As carbon emissions and energy consumption in computer systems always originate in computing time and effort, it is crucially important to find the most efficient ways to design computer programs to carry out a specific tasks within a certain quality margin.
Energy-efficient Data Centers:
Data centers are at the core of digital transformation. They host the ubiquitous cloud and execute millions of applications used by people all around the world. They enable people and institutions worldwide to access computing power and data storage on a massive scale and make all the popular digital applications possible, which accompany us in everyday life. However, data centers consume enormous amounts of energy. Coming to grips with this issue is mandatory to make digitalization sustainable in the future.
Clean AI:
Today, artificial intelligence has become a powerful tool to manage many tasks and to organize the modern connected world. Without AI we would not be able to manage the global flood of data collected by millions of digital devices. It would be impossible to navigate the vast WWW with its billions of websites and web documents. We could not make meaningful predictions from statistical data collected to diagnose diseases, organize traffic, retail and goods distribution. Without AI the vision of driverless cars and predictive maintenance would remain fantasy and none of the global sustainability challenges like fighting poverty, illiteracy, pollution etc. would be within reach. In short: A life without AI technologies has become hardly imaginable. However, the ever-growing demand for AI applications has its price. Training and executing the latest AI models based on deep neural networks require massive amounts of structured data and hundreds of thousands of computation layers. Training a powerful neural network can emit as much carbon as the life cycle of five cars including fuel. In order to have the benefits of AI without destroying the planet it is mandatory to find new training techniques and operating schemes for deep neural networks, which require much less data input, computation power and therefore much less energy.
clean-IT openXchange:
The clean-IT openXchange is a series of live talks and events on sustainable digitization. Once every month, experts present a topic on which participants can then directly ask questions and discuss. The topics of the talks consist of a mix of topics already presented in this forum and completely new ideas. The talks are open to everyone as online meetings. Participants are invited to exchange and discuss ideas with the presenters and viewers. In this section, the recordings of the talks will be uploaded subsequently.
