When comparing energy consumption patterns of AI infrastructure deployed in public and private cloud environments across different cloud architectures are there noticeable differences, comparing apple to apple? Can enough data be extracted from reliable and reputable sources to make such a conclusion?
To help with a detailed comparison, could you please clarify a few things:
Carl Beitsayadeh
Thank you kindly for your reply. 1) I am interested in comparing each workload separately, but if not possible, then the general mix would suffice; 2) I would like to include as many cloud systems as possible, and fairly equally for public and private; 3) I am interested in a global view, including diversified geographical locations where high usage is suspected: USA, EU, and asia-pacific regions; 4) beyond energy consumption, I am also interested in the workload efficiency. Additionally, if possible, I am interested in determining which operational factors (i.e., cooling efficiency, hardware specifications, or workload scheduling) contribute the most to the energy consumption & efficiency of both private & public clouds.
Muhammad, your insightful questions really stood out to me, and your educational background is impressive -- as a team of two faculty researchers from the University of Phoenix (USA) currently conducting a study on this very topic, we would be thrilled to have you join our research team in collaborate and delving deeper into this subject. Do let me know if you are interested and have the time to undertake such a task.
- Badges
- Science topic
- Similar topics
- Correction