Increasing Energy Efficiency of Server Cooling Using Deep Reinforcement Learning

Learn how Deep Reinforcement Learning (DRL) agents running on OCP-compliant BMC platforms can revolutionize server cooling efficiency in this technical presentation. Explore how Axiado's innovative approach moves beyond traditional thermal management techniques like static speed fans and PID controllers to create an intelligent cooling system. Discover how the DRL-powered dynamic thermal management agent learns and adapts to optimize temperature-energy balance while anticipating future cooling requirements based on workload patterns. Examine the practical benefits of this AI solution, including its ability to scale across diverse server environments with minimal human intervention, potential 40% reduction in fan usage, and 8% overall server energy savings. Understand the significant cost implications, with potential savings of €6m annually for 100K servers, making this solution particularly valuable for data center operations. The presentation demonstrates how a 0.5 TOPs tiny ML engine can transform server cooling efficiency while maintaining optimal performance.