Heavy Networking 633: Building DPU Apps With NVIDIA DOCA (Sponsored)

Heavy Networking - A podcast by Packet Pushers - Fridays

Categories:

Today’s Heavy Networking podcast dives into Data Processing Units, or DPU’s. More specifically, we’ll get under the covers of NVIDIA DOCA, a programming framework for NVIDIA Bluefied DPUs. The typical DPU has multiple ARM CPU cores, 4 or 8GB of memory, a crypto or acceleration engine, SerDes and Phy modules, but most importantly a networking ASIC. The DPU hardware is roughly similar to your everyday network appliance. But the most important aspect is the software that brings network and storage services inside the server. By using a set of  applications, libraries, and services, developers can build software to accelerate network, security, and storage functions directly on the DPU, freeing up the server CPU to focus on applications. How does this happen ? Its all software. In today’s sponsored show we explore DOCA on the Bluefield DPUs. DOCA is both a runtime operating system on the DPU including tools for provisioning, deploying, and orchestrating containerized services, but it’s also an SDK to supports a range of operating systems and distributions and includes drivers, libraries, and tools. Our guests are Justin Betz and Wes Kennedy, both Technical Marketing Engineers with NVIDIA. We discuss: * DPU essentials * The DOCA framework * DOCA applications, libraries, and services * Application use cases including firewalls, network overlays, storage acceleration, telemetry, encryption, and more * NVIDIA partners including VMware and Palo Alto Networks * More Show Links: NVIDIA DOCA Software Framework – NVIDIA Demystifying NVIDIA DOCA – NVIDIA Technical Blog DOCA SDK Documentation – NVIDIA