This is my research on attribution privacy, which I did for my final project in CSE450 Privacy Aware Data Analytics. tl;dr attribution privacy is when you have a P2P network where some peers are the original sources of the data, and they want to keep that private (they don't want the data to be attributed to them).
If you have pdflatex, ruby, and gnuplot you should be able to just type make
to build report.pdf. This generates all the data for the paper by running
thousands of network simulations, so you need a pretty fast system and a good
bit of RAM if you don't want it to take forever.
You can speed it up if you make the simulated networks smaller. They all use 100
peers currently, but you can change that by changing the *.data files use in
the *.gp plot files, and by changing the chart*.pdf inclusions in
report.tex. The Makefile has rules for running the correct network simulations
based on these file names.