Internet data centers, typically distributed across the world to provide timely and reliable service to local regions, are under increasingly greater pressure to cut energy costs and reduce their carbon footprint. As Internet data centers will soon be required to abide by carbon capping policies, there is a need for online framework that coordinates workload management between centers and optimizes joint energy consumption. The idea is to shift the workload toward greener data centers, but the problem is challenging due to the tradeoff between electricity cost and carbon footprint, as well as the inconsistency of renewable energy supply. Current solutions parameterize the carbon cap over long term operation of a data center (e.g. a year), which ignores the contiguous effects of renewable sources on carbon capping, and consequently, neglects to optimize energy costs relative to carbon emission allowance.
Researchers at ASU have developed OnlineCC, an online algorithm for minimizing operational cost while satisfying the carbon footprint reduction goals of a set of geo-distributed data centers (a cloud). The algorithm operates without needing future information, making it ideal for handling uncertainties such as electricity price, input workload, and renewable energy availability. Results show that OnlineCC reduces costs by 15-20% while maintaining an equal or smaller carbon footprint.
Potential Applications
- Carbon Neutrality
- Cloud Emissions Reduction
- Data Center Infrastructure Management
- Geo-Distributed Data Centers
Benefits and Advantages
- Comprehensive – Optimal solution for handling and predicting unknown factors such as electricity cost and renewable energy availability.
- Environmental – Cuts down global carbon footprint from internet data centers.
- Economical – Saves money by reducing energy costs by 15-20%.
For more information about the inventor(s) and their research, please see