Deep Learning Technology for Data Center Power Savings
We are looking for a deep learning technology to measure temperature data of 3-dimensional space, and can efficiently handle a large amount of data compression in high speed.
Why this is important
Data centers consume tremendous amount of power with both economic and environmental consequences. We are looking for a dynamic solution to optimize power use in data centers by prediction and distribution of computational load to maximize cooling efficiencies. Although technologies exist for optimizing cooling systems by dynamically adjusting fans, pumps, windows, etc. , to our knowledge, no systems exist to dynamically predict and distribute the heat generation created through computation to optimize cooling efficiency. For enhancement of data center power-saving by AI, grasping the temperature distribution of data center by thermography is required. Data compression of vast amounts of sensor data is also required. Though there are some candidate products or technology in Fujitsu Japan (FJJ), these are yet to be evaluated.
Description of the needWe are looking for technology that:
- Measures temperature data of 3-dimensional space: e.g. Expressing 3-dimensional temperature distribution on a 3-dimensionl model using a few 2-dimensional thermography.
- Efficient and fast data compression and extension of vast amounts of sensor data without lacking data.
- System designed for medium and larger data centers
- Initial target is 20% reduction in cooling costs.
- Higher levels of technology readiness preferred, but we are looking for technologies that have at least completed the research phase and are, at a minimum, demonstrable at the prototype level.
To be considered, submissions must include
- Contact information for the potential technology partner
- Description of the technology, its stage of development, and its capabilities
- Information on any patents granted or applied for (i.e., information in the public domain)
- Data demonstrating energy consumption estimation accuracy and efficiency improvements (if available)