Accelerating your network with RoCE for AI/ML and Storage fabrics

Register now for the 5 Nov webinar

Register
CalendarThursday, November 5 Time2:00pm AEDT PresenterMichael Schipp PresenterAshrut Ambastha

Today’s data centers demand that the underlying interconnect provide the utmost bandwidth with extremely low latency. RDMA over Converged Ethernet (RoCE) enables efficient data transfer on existing Ethernet infrastructure, without the expense of adding vast amounts of hardware or a large-scale conversion.

RoCE is critical to maximise server and network utilisation for AI and high performance storage applications such as Spark, TensorFlow, IBM Spectrum Scale, Lustre, Microsoft SMB, NetApp, PureStorage, DDN, VAST and NVMe-oF support in VMWare vSphere 7.0.

Join this upcoming webinar with NVIDIA's networking experts, as they discuss best practices, monitoring and fault finding with RoCE.

monitor-approveLearning Objectives

Come with a basic understanding of traditional interconnect protocols based on TCP/IP or UDP, and leave with advanced knowledge of:

  • The advantages of RoCE with storage
  • Understanding ECN and PFC in real application scenarios
  • VLANs and RoCE
  • RoCE for lossless fabric
  • Step-by-step guide to install AI clusters for NVIDIA DGX systems
  • Achieve advanced network virtualization with VMWare vSphere and RoCE