SNIAVideo Courses

SNIA Emerald Specification Updates and Related Programs

Discover SNIA Emerald V1.0 & V5.0 specifications for measuring enterprise storage device and system power consumption, regulatory compliance, and testing methodologies.

• YouTube
• 59 minutes
• Self-Paced
• Free Video

The Information Bank - Redefining Unstructured Data Storage and Management

Discover how computers' inability to understand stored data costs billions annually and explore a revolutionary two-pronged approach to intelligent storage systems.

• YouTube
• 54 minutes
• Self-Paced
• Free Video

Why NVMe Needs to Evolve for Efficient Storage Access from GPUs

Explore how NVMe storage protocol must evolve to efficiently handle GPU-based AI workloads, comparing CPU vs GPU execution models and identifying key bottlenecks.

• YouTube
• 49 minutes
• Self-Paced
• Free Video

Highly Scalable, Masterless, Distributed File System at Rubrik

Explore Rubrik's masterless distributed filesystem architecture, covering erasure coding, data integrity, performance optimization, and challenges in cybersecurity environments.

• YouTube
• 50 minutes
• Self-Paced
• Free Video

SSD Virtualization - Enabling New Possibilities for SSD-Based Ecosystems

Explore revolutionary NVMe and OCP features transforming SSD virtualization, covering SR-IOV, VM migration, QoS management, and security extensions for robust virtualized environments.

• YouTube
• 51 minutes
• Self-Paced
• Free Video

Global Distributed Client-side Cache for HPC and AI Workloads

Explore distributed client-side caching solutions for HPC/AI workloads, addressing communication bottlenecks and leveraging aggregate cluster resources for microsecond latencies.

• YouTube
• 51 minutes
• Self-Paced
• Free Video

FAMFS - Get Ready for Big Pools of Disaggregated Shared Memory

Explore CXL-enabled disaggregated shared memory and the open-source FAMFS file system for memory-mappable files with minimal app modifications.

• YouTube
• 51 minutes
• Self-Paced
• Free Video

Choosing Your AI Storage Protocol - A Deep Dive into SMB and NFS Performance, Tuning, and Overhead

Explore SMB vs NFS protocols for AI workloads through MLPerf benchmarks, comparing performance, configuration complexity, and operational overhead to optimize storage decisions.

• YouTube
• 48 minutes
• Self-Paced
• Free Video

Model Training in Public Clouds - Case for IBM Storage Scale

Explore IBM's Cloud Vela cluster architecture using distributed file systems over object storage for AI model training in public clouds with Kubernetes orchestration.

• YouTube
• 48 minutes
• Self-Paced
• Free Video

Disaggregated KV Storage - A New Tier for Efficient Scalable LLM Inference

Discover how disaggregated key-value storage architecture reduces GPU memory pressure and achieves 5-8× higher throughput for scalable LLM inference systems.

• YouTube
• 55 minutes
• Self-Paced
• Free Video

Assessing AI Storage Communication Performance at Scale

Explore methodologies for assessing AI network and storage infrastructure performance without expensive GPU provisioning, covering topologies, benchmarking, and optimization metrics.

• YouTube
• 44 minutes
• Self-Paced
• Free Video

Towards Memory Efficient RAG Pipelines with CXL Technology

Explore CXL memory pooling and tiering to optimize RAG pipeline efficiency, reduce DRAM costs, and handle dynamic memory demands in AI inference with minimal performance impact.

• YouTube
• 34 minutes
• Self-Paced
• Free Video

Storage for AI 102

Explore advanced AI workflow phases and their distinct storage access patterns, plus guidelines for optimizing storage systems for different AI processes and workloads.

• YouTube
• 50 minutes
• Self-Paced
• Free Video

Scaling RAG with NVMe

Explore DISKANN's hybrid approach to scale vector databases using NVMe SSDs, reducing memory footprints while maintaining query performance for petabyte-scale RAG applications.

• YouTube
• 49 minutes
• Self-Paced
• Free Video

Towards Unified Knowledge Platforms - Evolving Storage Systems for Generative and Agentic AI

Explore how storage systems evolve into unified knowledge platforms supporting Generative and Agentic AI workflows with real-time metadata enrichment and vector embeddings.

• YouTube
• 40 minutes
• Self-Paced
• Free Video