A test suite for QUIC

Note: This project has moved to https://github.com/QUIC-Tracker

This work is part of my master thesis for my master's degree in Computer Science at Université catholique de Louvain. It is supervised by Pr. Olivier Bonaventure and Quentin De Coninck.

It is two-fold:

• It is a test suite that exchanges packets with IETF-QUIC implementations to verify whether an implementation conforms with the IETF specification of QUIC. The test suite is consisting of several test scenarii. Each of them aims at testing a particular feature of the QUIC protocol. The test suite runs daily, and its results are available on https://quic-tracker.info.ucl.ac.be. This is the main focus of my effort.

• It also contains a tool that collects HTTP Alt-Svc headers sent by web-servers of popular domain names in the hope of discovering IETF-QUIC capable hosts. It lists the versions advertised by the hosts, if any.

Installation

Test suite

The test suite comprises a minimal Go implementation of QUIC which is currently draft-13 and TLS-1.3-draft-28 compatible, as well as several test scenarii built upon this implementation. The test suite outputs its result as JSON files, which contains the result, the decrypted packets exchanged, as well as a pcap file and exporter secrets.

You should have Go 1.9, tcpdump, libpcap libraries and header installed before starting.

go get github.com/mpiraux/master-thesis
cd $GOPATH/src/github.com/mpiraux/pigotls
make

The test suite is run by bin/scenario/scenario_runner.go. For help about its usage see:

go run bin/scenario/scenario_runner.go -h

It takes as first parameter a tab-delimited CSV containing a list of hosts and particular URLs which will generate data when requested.

A second optional parameter exists to specify a particular scenario from the directory scenarii to run exclusively, e.g. stream_opening_reordering.

The results are printed to stdout.

Web application

The web application is a Python Flask application that presents the test results in an human-readable way.

It is known to be working with Python 3.6, but it should be compatible with earlier Python 3 versions. It requires the following packages:

pip3 install flask pyyaml sqlobject

Then

• Fetch web dependencies using yarn install in quic_tracker/static

• Add the project root directory to $PYTHONPATH using export PYTHONPATH=$PYTHONPATH:$PWD

• Start the application with python3 quic_tracker/app.py

• Output from the scenario runner should be placed into quic_tracker/traces with a name in the format \d*.json