Skip to content

lbarman/prifi-experiments

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

8 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

PriFi Experimental Data & Plots

This repository contains all raw data (and their derived plots) about experiments and benchmarks of PriFi. The goal is to have reproducible plots from raw data.

Prerequisites

  • Python2
  • NodeJS + npm
  • underscore JS (just get all dependencies via npm install)
  • Gnuplot
  • epstopdf (typically comes with the LaTeX full suite)

Structure of this repo

Please ignore the root folder! It only contains a bunch of scripts used by the sub-folders.

Go to, e.g., fig3a-e2e-latency (the numbering of the figures follow the paper) and examinate the Makefile to understand what is going on.

Structure of each subfolder:

  • config : a copy of the sda/simul folder (in PriFi) at the time of the experiment; it contains all parameters
  • logs : raw logs exported from the relay. Raw logs are .txt files

Most Makefiles do the same thing:

  1. parse the raw log into JSON format
  2. extract the relevant lines (e.g., latency) into a separate files (e.g., measured-latency.compiled). Note: this is just a fancy grep, and the output is text
  3. compute statistics (often: simply the mean of the value after the steady-state occurs, excluding obvious failures such as runs with timeouts, etc). This runs the aggregate.js main script with a .json (data source) and a .js (processor script). Creates a yyy.gnudata file that can be plotted
  4. eps (runs gnuplot using the xxx.gnuplot script and the yyy.gnudata formatted input)
  5. pdf (converts eps to pdf)

Author: Ludovic Barman