ACETA: Accelerating Encrypted Traffic Analytics on Network Edge

In Summer 2019, I was fortunate enough to take part in a 10-week research program with my mentor, Dr. Peilong Li, as a member of the ACETA team. This project's goal was to design an encrypted traffic analytics system with accelerated speed to be implemented on x86 multicore platforms that are common on most enterprises' network edge. My role in this project involved implementing Intel performance libraries for machine learning models, and updating the legacy Python 2 code I inherited to Python 3. The two Intel libraries implemented were the Intel Data Analytics Acceleration Library (DAAL) for several machine learning models and the Intel OpenVINO Toolkit which was used to optimize a trained Keras-Tensorflow neural network. At the conclusion of the research period, a paper detailing this project and its results was accepted to IEEE's International Conference on Communications, which I presented remotely to the conference.

Project Paper