Machine Learning Inference Support in DPDK for Network Applications
Watch a 22-minute conference talk exploring how Machine Learning inference capabilities are being integrated into DPDK (Data Plane Development Kit). Learn about the process of generating predictions using pre-trained models for emerging networking applications, including channel estimation in next-gen radio networks, anomaly detection, and network traffic analysis. Discover how the MLDEV specification enables low-latency inference operations directly in the datapath while utilizing packet buffers for computation. Explore recent developments in ML inference hardware such as integrated accelerators, embedded devices, and FPGAs, and understand how DPDK's software framework can leverage these technologies for inline inferencing. Follow along with a detailed walkthrough of the lib mldev library and gain insights into the future roadmap of MLDEV implementation.