ABSTRACT

While the growth of cloud platforms has fueled the rise of a diverse set of online services in recent decades, it has also led to increasing energy consumption and, hence, carbon emissions. Cloud platforms are well-positioned to reduce their carbon emissions by transitioning to cleaner energy sources because cloud applications often have significant spatial, temporal, and performance flexibility, enabling them to shift the location, time, and intensity of their execution to better align with the availability of carbon-free renewable energy or low-carbon grid energy. Unfortunately, researchers cannot leverage this unique combination of advantages to experiment with and optimize cloud applications' carbon-efficiency because current cloud platforms do not expose energy's carbon characteristics to them. To address the problem, this project will design and implement a shared community testbed for experimenting with the design of carbon-efficient cloud applications. The testbed will be deployed in an edge data center with a local energy system that includes a co-located solar array, batteries, and cooling system. The testbed software will virtualize the energy system by exposing software-defined visibility and control of it to cloud applications, which will enable experimentation with a rich, but unexplored, design space for developing novel carbon-efficient cloud applications capable of responding to clean energy and carbon dynamics.

The project has the potential for significant technical impact in advancing the design of carbon-efficient cloud applications, which is important for reducing environmental damage associated with cloud platforms' increasing energy usage and carbon footprint. The project will involve significant community outreach, including annual workshops and tutorials, to both raise awareness of the importance of optimizing for carbon-efficiency and demonstrate how the testbed can enable and advance carbon-efficiency research. The project also plans to conduct outreach by incorporating sustainable computing topics as part of summer programs for local middle and high school students. The project will incorporate use of the testbed into current courses on cloud computing and green computing to both get feedback on the testbed and to enrich students' learning environment. Finally, in addition to making the testbed available to the community, the project will make the software artifacts and datasets developed by the project available to the community via the UMass Trace Repository, which hosts many open datasets collected by researchers.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

Please report errors in award information by writing to: awardsearch@nsf.gov.