In this blog series, we are presenting the five P4 projects accepted for this year’s Google Summer of Code (GSoC). The fifth one is Integrating P4 Based In Network Machine Learning Framework into P4Pi.
GSoC Project
The fifth one is Integrating P4 Based In Network Machine Learning Framework into P4Pi
This project integrates Planter, an in-network machine learning framework, into the P4Pi Raspberry Pi platform. The work includes adding a p4c-dpdk target adapter to Planter, implementing 1–2 additional ML algorithm modules, developing end-to-end example applications on a DPDK-based software switch, and packaging the result as a ready-to-use P4Pi system image with teaching-oriented documentation and tutorials.
This project is to improve the P4Sim module by enhancing its Traffic Manager and integrating ns-3’s built-in QueueDisc class into it. This will help developers and researchers in testing P4 switches with more complex scenarios and get accurate results for their experiments.
Project Mentor: Peng Qian
Contributor: Yuzhong (WeiWei) Luo
Yuzhong is a first-year MEng Engineering Science student at Trinity College, Oxford, with a background in competitive programming (USACO Platinum) and ML systems research. I work as an undergraduate researcher at Oxford’s Visual Geometry Group and will be integrating the Planter in-network ML framework into the P4Pi platform this summer.
Connect with Yuzhong on LinkedIn
