MCPA (Mobile Critical Path Analysis)

Open dataset of MCPA paper "Tackling Mobile Traffic Critical Path Analysis With Passive and Active Measurements", to appear in TMA 2019 conference.

Contents

• performance_metrics.ipynb
  A Jupyter notebook to display/plot/compare performance indicators (e.g., Page Load Time, Above-the-Fold Time, Trasport Delivery Time, ..) for the mobile apps and webpages used in the paper. Based on the logs in ./performance_metrics.

• critical_path_info.ipynb
  A Jupyter notebook summarizing critical path information for the mobile apps used in the paper. Based on data in ./critical_path_info.

• waterfall.ipynb
  A Jupyter notebook generating network waterfalls for mobile apps used in the paper. It uses the logs in ./waterfalls.

• ./tool
  Part of MCPA code for the processing of pcap mobile-traffic traces. It partitions traffic into activity windows. For each interval, it computes performance metric(s) and reports statistics on the network waterfall and on the Critical Path.

Getting started with MCPA tool

Requirements

• Python 2.7+
• Haralyzer (https://pypi.org/project/haralyzer/), Harparser (https://pypi.org/project/harparser/)
• NumPy

Example (Facebook, app-startup traffic)

cd tool/
python main.py -p ./examples_app-startup/com.facebook.katana.pcap -c critical_domains_app-startup.txt 

Arguments

-p : Input pcap file.
Examples can be found in ./tool/examples_app-startup for app-startup traffic and ./tool/examples_app-click for app-click traffic

-c : File containing the set of critical domains of the mobile app.
Use ./tool/critical_domains_app-startup.txt for app-startup traffic and ./tool/critical_domains_app-click.txt for app-click traffic