Unless you live a completely tech-free life, use a computer, laptop, or other type of equipment. Unbeknownst to you, you are in the vast universe called Green Computing. The concept is defined in specialized publications, Green Computing representing the responsible and ecological use of computers and their resources. More broadly, it is also defined as the study of designing, engineering, manufacturing, using and recycling computing devices in a way that reduces their impact on the environment. Companies that use green computing methods, such as Supermicro, often deploy energy-efficient central processing units (CPUs), servers, peripherals, power systems, and other IT equipment. Supermicro’s strategy focuses on reducing the use of resources, in parallel with improving the performance of the equipment, the innovation ensuring the company’s leading position in Green Computing
Information technology in kWh
If you are reading this article online, in your computer or mobile phone there is a transaction between the memory and the processor, which involves the consumption of energy. As these tasks become more elaborate and data-intensive, two things begin to grow exponentially: the need for more memory and more energy consumed. The internet and technology are advancing every day, increasing the number of new devices coming to the market at the same time. At the same time, the computing power of a device is increasing, so is the power consumption of these devices.
The numbers are impressive! According to Digital Information World, information and communication technology, including data centers, communication networks and user devices, accounted for between 4 and 6% of global electricity consumption in 2020. If we continue at this rate, by 2030 it is expected to increase by 8-21%. It is estimated that this consumption is relatively equally divided between users (55%) and producers (45%).
The aforementioned study shows that the Internet consumes 70 billion kWh/year, Google consumes about 0.013% of the world’s energy, accessing Amazon consumes 0.0003 kWh, and watching YouTube videos for 5 minutes requires 0.065 kWh. There are approximately 509,147 data centers worldwide, and their associated carbon emissions doubled between 2017 and 2020, with each facility consuming between 20 and 40 megawatts, enough to power approximately 16,000 households.
Let’s not forget that Artificial Intelligence (AI) has been heavily promoted by companies due to its ability to improve energy efficiency in many industries. According to a McKinsey report, AI has the potential to deliver energy savings of up to 20% in buildings and 15% in transportation systems. In addition, AI-based solutions can help companies reduce CO2 emissions by up to 10% and reduce energy costs by 10-20%. However, some real costs of using this technology cannot be neglected, and the power consumption inherent in using AI should be a primary concern. According to TechTarget, the total consumption of a program that analyzes data sets to create models and make predictions over nine days was no less than 27,648 kWh, the amount of energy used by three households in a whole year.
These are just some data that show how much energy the technology sector consumes. With the development of new devices, energy consumption will also increase rapidly. Although there are efficient ways to produce electricity today, the resources may not be sufficient to meet the future energy needs of technology centers.
Green Computing, a solution
The time has come to pay attention to the concept of Green Computing, which emphasizes the efficient and correct use of computers and their computing processes. Innovation is now turning to green computing, managing computing systems to reuse, recycle and reduce negative impact on the atmosphere.
Green Computing, also known as Green IT or sustainable computing, emerged in response to growing concern about the impact of information technology on the environment and the rapid proliferation of electronic devices and data centers. The earliest discussions of the environmental impact of computers date back to the 1970s, with concerns about the energy consumption of large mainframe computers. Researchers have begun to study the effects of computing on the environment, although these discussions have remained relatively niche.
In 1992, the United States Environmental Protection Agency (EPA) launched the Energy Star Program, considered the starting point in Green Computing. Products that met Energy Star standards were labeled as energy efficient. Between 2000 and 2010, data center operators began implementing energy-efficient practices such as hot-cold aisle isolation, free cooling, and server virtualization to reduce energy consumption and cooling costs.
Green Computing continues to evolve with innovations in energy-efficient hardware, cloud computing and sustainable software solutions as a holistic approach to reduce the carbon footprint of information technology, with the aim of balancing the benefits of technology with the need to protect the environment and to conserve resources.
Supermicro Solutions. Case Study
Green eDIH ecosystem partners, whether from hardware or software, have turned their attention to Green Computing. Global Systems Communications (https://gsc.com.ro/), a founding member of Green eDIH, has extensive experience in the design and integration of solutions based on state-of-the-art technologies, from industrial automation to high performance and efficiency servers , developing long-term partnerships for each project, both with customers and solution providers. One of the strategic partners, for which GSC is an authorized distributor and integrator, is Supermicro (https://www.supermicro.com), a global leader in high-performance and efficient servers, storage technology and Green Computing for HPC, Data Center, Cloud Computing, Big Data and Embedded Systems.
Supermicro has demonstrated that data centers can be more energy efficient with energy saving solutions, liquid cooling systems and reuse of existing equipment to improve a data center’s carbon footprint while reducing power usage efficiency (PUE).
Supermicro servers are designed to require less power at a given performance level. This feature is achieved through a design that shares components such as fans and power supplies, which can reduce power consumption by approximately 10%. In addition, individual subsystems such as the processor, memory or storage components are designed to be upgraded without the need to replace the entire chassis. Using this disaggregated design greatly reduces e-waste and costs when purchasing new technology.
Traditional cooling systems required massive amounts of water and were a huge expense. Conventional methods of using water for heat exchange, together with cold water cooling plants, are not only expensive, but also put a strain on natural resources. Supermicro systems are designed for air cooling, with servers designed for maximum airflow, resulting in the processor’s ability to operate at higher temperatures.
Supermicro invests heavily in the design of its equipment, which provides maximum performance and reduces power consumption. Designing a server involves serious knowledge of electronics, mechanical design, airflow simulation, thermal dynamics, and packaging. The company uses advanced computational fluid dynamics (CFD) simulations, including complex algorithms that simulate the paths of air molecules over heat-sensitive hardware components, to optimize server performance and efficiency.
Green Computing is the Supermicro standard in server design. This resulted in the top spot in the 2020 Green500 ranking, when the MN-3 supercomputer, built in collaboration with Preferred Networks, achieved a record 21.11 Gigaflops of performance-per-watt in a benchmark run which delivered a total performance of 1.62 Petaflops. This efficiency achievement is 15% higher than the previous Green500 record of 18,404 Gflops/W, which was set in June 2018.
The company believes that if more organizations applied Green Computing principles, they could save a combined cost of 10 billions of USD in electricity per year. Reducing electricity consumption using Green Computing solutions like Supermicro is equivalent to eliminating 30 fossil fuel power plants and that is also equivalent to planting approximately 8 billion trees to offset carbon emissions.
More details about best practices in Green Computing can be found here:
