Nathan Tracy, OIF President, underscores the rising importance of machine learning and artificial intelligence in data centers, driving the need for increased bandwidth and density.
- The future data centers will require a significant shift in bandwidth and bandwidth density to meet the demands of machine learning and artificial intelligence. This will inevitably drive power consumption and thermal management in these data centers.
- OIF is working on several key developments to enable the next generation of hyperscale data centers. These include the 200 Gbps electrical interface developments, which are optimized for the lowest possible power, and the development of the 112 Gbps Linear Optics specification, which provides significant savings on power consumption and cost.
- OIF has also initiated the Energy Efficient Interfaces project, which takes a holistic view of all the electrooptical link requirements needed in an AI data center. This includes fabric data rates and latencies required directly at the GPU and CPU interface, as well as the backend fabric where 200 gigabit will be a critical building block for next-generation bandwidth and bandwidth density.
