AI Proving GroundCisco Data Center NetworkingAI SolutionsATCAI & DataCisco
Learning path

Cisco RoCE for AI/ML

Skill Level
Fundamentals
Duration 2 hours 50 minutes
Updated Nov 20, 2024

About this learning path

The user of this learning path will learn the components of RoCE and why it is essential for clean, fast, and reliable AI/ML compute communication.

Your instructors

Craig KemmererWorld Wide TechnologyTechnical Solutions Architect
Michael WitteWorld Wide TechnologyPrincipal Solutions Architect

Prerequisites

  1. CCNA certification or a basic understanding of primary network concepts, VLANs, IP addresses, and Gateways.
  2. Having a conceptual knowledge of network fabric and fabric design is preferable.

What you'll learn

  1. RDMA Over Converged Ethernet Version 2 (RoCEv2)
  2. Priority Flow Control (PFC)
  3. Explicit Congestion Notification (ECN)
  4. Data Center Quantized Congestion Notification (DCQCN)

Technologies

Learning path
Collapse all
Cisco RoCE for AI/ML
  1. 1. The theory behind RDMA over Converged Ethernet (RoCE)
    1. Enroll in this learning path to view locked contentUsing PFC and ECN queuing methods to create lossless fabrics for AI/ML
      Article
      Locked
    2. Enroll in this learning path to view locked contentUnderstanding Data Center Quantized Congestion Notification (DCQCN)
      Article
      Locked
    3. Enroll in this learning path to view locked contentIntroduction to Cisco's AI/ML GPU Networking Solution
      Article
      Locked
  2. 2. Practical use of RoCE
    1. Enroll in this learning path to view locked contentCisco RoCE Fabrics
      Lab
      Locked
  3. 3. Conclusion
    1. Enroll in this learning path to view locked contentQuiz
      Quiz
      Locked
    2. Enroll in this learning path to view locked contentLearning Path Complete
      Achievement Badge
      Locked