As Moore's Law comes to an end and data-center computing becomes more dominant, the importance of domain-specific hardware-software co-design and full-stack integration is increasing. This is crucial to achieve better performance within the given total cost of ownership (TCO). In this presentation, I will first discuss how FADU has achieved high performance and low power in the data center-grade NVMe SSD domain through its unique hardware-software co-design. I will also provide an overview of how the domain-specific hardware-software co-design technique and design principles applied to the SSD domain can be expanded to various computer systems. Additionally, I will discuss the challenges that major hyperscalers such as Amazon, Microsoft, Google, and Meta are facing, as well as the opportunities available. In addition to presenting technological and product perspectives, I will also share my experiences in building and scaling teams and achieving success in the competitive data center industry.

Dr. Eyee Hyun Nam is the co-founder, CEO, and CTO of FADU Inc., which was established in 2015. The company focuses on developing system-on-chip (SoC) solutions for data centers, including SSD controllers, CXL switch, and Power management chips. Dr. Nam is responsible for leading the entire development and productization process, which encompasses SoC, firmware, validation, and mass production. Dr. Nam earned his B.S. degree in Electrical Engineering and his Ph.D. degree in Computer Science & Engineering from Seoul National University in 1998 and 2011, respectively. During his post-doctoral research at SNU for 2 years, he concentrated on SSD reliability, parallelized FTL, and compositional constructions of flash storage systems. Following the establishment of FADU Inc., Dr. Nam has successfully developed and commercialized multiple generations of NVMe SSD products, renowned for their low power consumption and high performance. His research and development interests span application-specific hardware/software co-design, co-optimization of large-scale computing systems for enhanced performance and total cost of ownership (TCO), full-stack memory and storage system design and the development of programmable hardware accelerators for data center services.