There are many excellent reasons to move your analytical workloads and data storage to the cloud, and for many users and organizations, sustainability has become a primary driver for that move. These concerns are even more prominent in research, as funding organizations mandate sustainability as a part of the grant application process.
One example in the US is the National Science Foundation (NSF), whose Proposal and Awards Policies and Procedures Guide state: “(p)rior to making an award, NSF considers the anticipated environmental impacts associated with the activities described in the proposal.” The UK’s largest research funding organization, United Kingdom Research and Innovation (UKRI), which grants over £7 billion a year, has pledged to “embed sustainability in everything we do” and has challenged researchers to think “how can we meet our environmental obligations and still foster excellence and impact in research and innovation?”. Meanwhile in Canada, the National Science and Engineering Research Council (NSERC), the largest of the three main research funding organizations in Canada, which had a $1.7 billion CAD budget in 2022 has a stated goal - as part of the Canadian government’s broader Greening Government strategy - of “aid(ing) the transition to a net-zero, circular economy through green procurement that includes lifecycle assessment principles and the adoption of clean technologies and green products and services”.
Note that these sustainability goals and pledges are not just for environmental research, they apply to *all* research funded by these agencies across every discipline of science, engineering, medical, and social science. Fortunately, cloud computing provides multiple ways to reduce the environmental impacts of research and analytical computing. Without further ado, here are the top 5 ways the cloud directly supports sustainability in research:
The UKRI challenges researchers with the following questions:
“If you’re studying the deep ocean, for example, how do you think about environmental sustainability in that context? Are there ways to explore the ocean as we do the stars? Are there non-carbon ways of using ships – or other technologies we can use instead? How can we travel less in undertaking research? In other words, how can we meet our environmental obligations and still foster excellence and impact in research and innovation?”
While field research will never disappear, there are opportunities for novel research with lower impact. For instance, at Oracle for Research we are working with a research team monitoring agricultural yield in Nigeria. They start with satellite data, use machine learning to identify field boundaries and map crop distributions at different times of the year, and analyze where there may be improvements to be made in type of crops planted, field sizes, or locations. Advances in remote sensing, collaborative analytical platforms, and artificial intelligence have therefore enabled remote work and alleviated the need for travel in research, resulting in greater cost efficiency and fewer transport emissions.
Similarly, we are working with biomedical researchers who can develop candidate drugs to treat specific diseases using three-dimensional molecular modeling and protein engineering. Cloud resources meant that a recent model run was able to simulate over 650,000 atoms over 500,000 steps using 8 NVIDIA A100 tensor core GPUs and 64 CPUs. Advances like these in molecular dynamics have not only drastically reduced the time to develop and create new medicines, they also eliminate the environmental effects of months (or years) of wet lab work which can now be simulated accurately.
One the largest environmental impacts of research collaboration is the travel required to meet and work with collaborators from different parts of the world. The global COVID-19 pandemic showed that communication and collaboration can continue when in-person contact is not possible, and this has led to a move to reduce, where possible, the volume of travel required.
Cloud data storage and virtual server resources facilitate effective research collaboration. Oracle for Research is supporting a project where researchers from Sri Lanka, Italy, and the US are working together on a single data set available to all participants. In fact, such collaboration can be quicker and more effective since different people access the data at different times of the day. The ability to access computing resources via SSH, VNC, or HTTPS from anywhere in the world with internet connectivity allows for much greater collaboration with minimal environmental impact.
Efficiency of utilization is one of the biggest arguments for using cloud infrastructure – from both economic and sustainability viewpoints. Shared resources on campus or in a regional consortium share some of these advantages (see below) but the hyperscale cloud vendors have made a science of optimizing resource usage as it directly impacts their costs.
Oracle for Research works closely with our resource planning colleagues to ensure that capacity that is not being used for commercial workloads can be used effectively for research analyses. Because it’s simple to move data around the world efficiently, it is quite common for researchers in Brazil to run workloads in our Kyoto region when capacity is available there (data sovereignty rules permitting). While there is a small energy cost associated with moving a few terrabytes of data around the world, this is offset by the efficient use of otherwise idle resources in a remote data center.
In parallel with dense deployment and high utilization, energy efficiency is designed into our hardware from the beginning. Cloud data center hardware uses 38% less power than commodity computing and storage infrastructure. Efficiency in power management, cooling, distribution, and other factors mean that cloud computing can reduce overall energy requirements for computing by 93%.
Of course, this has less impact if the power used in these data centers comes from polluting sources. Oracle Cloud is already powered 100% in Europe by renewable energy and we are committed to 100% renewable energy use across all regions by 2025. A combination of a commitment to renewable energy resourcing and efficiencies leads to a 98% reduction in CO2 emissions for cloud computing compared to traditional on-premises deployments.
The final consideration for sustainability in this context is the circular economy – the re-use and recycling of resources. Cloud computing already contributes significantly to this goal in that virtual resources in the cloud can be assigned and reassigned quickly and easily as customers’ needs change. Researchers don’t need to purchase a server that will be used for a project, and then need to be repurposed or disposed of; they can provision the computing and storage needs for any project virtually and then return them to the pool to be reassigned on completion.
At the hardware level, Oracle has designed our cloud equipment to be recyclable and reusable. In FY22, Oracle collected 2.8 million pounds of retired hardware assets, of which 99.9% was either reused or recycled.
Sustainability doesn’t just affect our research on environmental issues and public health, it is an important part of every research project. Concerns about environmental impacts are becoming a more central concern for people everywhere and Oracle is committed to doing our part for a cleaner and brighter future.
Please contact us at Oracle for Research if you’d like to know more about our support for sustainable research, or how OCI can help lower the environmental impacts of your research efforts.
Oracle for Research offers products, capabilities, and resources purpose-built for research to simplify the research process and accelerate discovery around humanities’ most urgent needs. Through our research-first, partner-centric approach, we make computing paradigms more powerful and user-friendly, data easy and accessible, and community resources invaluable and actionable. We help researchers in academic, commercial, and federal settings, across all research disciplines, explore novel ways to achieve ground-breaking results to make the world a better place.
Communication is Andrew’s first passion, having had his academic research and writing published internationally on multiple occasions. Andrew developed his marketing acumen at Switch Automation, Alveo and CACI International, where he delivered a range of technical sales content in a variety of formats and planned, project managed and executed on diverse marketing campaigns. Andrew has nurtured a creative outlook, having edited corporate videos at award-winning film studios in London and produced, shot and edited videos at Switch.