Google, IBM & Meta Certificates – 40% Off
One plan covers every Professional Certificate on Coursera.
Unlock All Certificates
Explore a fundamental rethinking of Jim Gray's legendary 5-minute rule for data placement in this conference talk that challenges decades-old assumptions about memory and storage trade-offs. Learn how the original 1985 rule, which balanced memory cost against disk access savings, requires updating for modern computing realities where infinite DRAM bandwidth and negligible CPU-GPU I/O costs no longer apply. Discover a new analytical model grounded in real hardware constraints that addresses the limitations of traditional approaches. Examine how this framework applies to contemporary AI-centric workloads including inference, graph neural networks, and large-scale data analytics where massive working sets frequently exceed DRAM capacity. Understand the implications for the Storage Next initiative's goal of leveraging NAND flash as an active tier for GPU workloads. Gain quantitative foundations for evaluating memory versus storage trade-offs and making informed caching decisions in next-generation computing systems through insights from industry experts at ScaleFlux and NVIDIA.
