Explore a 41-minute keynote presentation that delves into the significance of Compute Express Link (CXL) technology for the storage community, delivered by Dr. Yang Seok Ki from Samsung Electronics. Learn about the evolving landscape of data center architectures driven by exponential growth in data volume and traffic, focusing on the disaggregation and composability of computing resources. Gain insights into CXL's role as a key enabling technology for high-speed, low-latency interconnects with memory cache coherency between various computing components. Discover Samsung's CXL memory technologies, including memory expanders and memory semantic SSDs, while understanding the technical aspects of memory interfaces, cache coherence, access latency, and memory capacity challenges. Master the concepts of hardware abstraction, basic software implementation for memory expansion, and the convergence of memory expansion fabric into CXL. Examine the economic considerations of memory emulation using storage, extended memory subsystem architectures, and the emerging field of computational storage/memory devices.
Understanding CXL Technology: Memory Expansion and Storage Integration in Modern Data Centers
Overview
Syllabus
Intro
Technological Momentum
New Business Opportunities
Additional Technical Definition of CXL
CXL from the Memory Interface Perspective
CXL from the Cache Coherence Perspective
CXL from the Access Latency Perspective
Memory Capacity and Bandwidth Challenge
Convergence of Memory Expansion Fabric into CXL
Samsung Memory Expander
Hardware Abstraction for Memory Expander
Basic Software for Memory Expander
Economical Memory Challenge •Memory emulation using storage with large page movement through heavy software stack for better TCO
Extended Memory Subsystem Architectures
Hardware Abstraction for Memory Semantic SSD
Basic Software for Memory Semantic SSD
Computational Storage/Memory Device Concept
Taught by
SNIAVideo
