Paper Title: Hybrid Cloud and AI Through the Lens of 3D-NAND Technology

Paper Abstract: AI-centric workloads are the fastest growing applications driving the Hybrid Cloud infrastructure growth. As cloud IaaS (Infrastructure as a service) adoption is growing, flash-driven performance storage services will be essential. Snapshot and restore functionality as enabled by flash is becoming a key market differentiator for mission critical enterprise workloads. Flash storage also contributes towards sustainability and reduction in carbon footprint by using less power than HDDs. All these factors together result in significant reductions in TCO. In the last decade, flash-based storage systems have become mainstream offerings. This is primarily due to 3D NAND scaling, especially QLC, which allows further flash bit cost reduction, higher density, low latency, high throughput, high IOPS and overall lower TCO. This presentation will explore flash technology innovations focused on >200+ layers scaling, TLC/QLC/PLC adoption, and NVMe and PCIe Gen 4/5. It will also address 3D-NAND power/thermal issues, and the overall effect of higher temperature on drive and system performance, TCO, and its relationship with system quality and reliability.

Paper Author: Ranjana Godse, Non-Volatile Memory Engineer, International business machine (IBM)

Author Bio: Ranjana is an Advisor Engineer in IBM's Systems Supply Chain Organization, focusing on Semiconductor Technologies used across all IBM Systems and products. Ranjana is responsible for Advanced Memory Technology like 3D NAND and Storage Class Memory (SCM) Qualification and for IBM Server & Storage applications. She holds a MS in Electrical Engineering from Rochester Institute of Technology (RIT) and a B-Tech from Govt. College of Engineering, India.