Composable Memory Fabric Architecture Enabling Scalable Memory Management for AI and General Computing

Learn about composable memory fabric architecture designed to address the growing memory demands of AI workloads, in-memory databases, and caching applications in this technical conference talk. Discover how Cache-coherent interconnects like Compute Express Link (CXL) enable dynamic memory resource provisioning tailored to specific workload requirements. Explore the OCP CMS Memory Fabric Orchestration Workstream's collaborative approach to developing open and interoperable memory orchestration solutions through partnerships with standards organizations and ecosystem vendor reference implementations. Examine the detailed memory fabric architecture that leverages CXL and DMTF interfaces to enable data center orchestration tools like Kubernetes to efficiently manage memory resources across entire data centers. Gain insights into the future roadmap and focus areas for the CMS Memory Fabric Orchestration Workstream as presented by electrical engineer Hongjian Fan from Seagate and senior principal engineer Anjaneya "Reddy" Chagam from Intel.