This git is a decentralized simulation of IOTA which is a crypto-currency using OMNeT++. It implements two kind of modules : the malicious ones that can launch an attack and the honest ones representing the regular user.

This simulation can use several topologies:

• Complete graph
• 2D grid
• Torus
• Watts-Strogatz

As the simulation is handling two types of modules and the .NED format is very restrictive, you have to use the topologies folder which contains the GenTopo.py script allowing to generate the wanted topology. The simulation will then read the .csv created by the script to build the network (this part is done by the ConfiguratorModule). It is also necessary to install the networkx and the matplotlib modules.

The simulation implements three different TSA (Tips Selection Algorithm):

• IOTA: www.descryptions.com/Iota.pdf
• G-IOTA: https://ieeexplore.ieee.org/document/8845163
• E-IOTA: https://ieeexplore.ieee.org/document/9223294

And two types of attacks explained in the IOTA paper:

• Splitting Attack
• Parasite Chain Attack

The data folder contains all the tracking files allowing to:

• generate an image of the Tangle using the script TangleGen.py (you need however to install the graphviz module). The SVG figure will then be saved in the image folder.
• compute an average of the number of tips at the end of each simulation and the simulation execution time using the Stats.py script (glob module needed).

Moreover, in order to evaluate the success of each of these attacks, the simulation can also compute and export the following data in the tracking folder:

• the percentage difference in size between the two branches (splitting attack)
• the number of transactions that are part of the parasite chain (parasite chain attack)

In order to achieve an acceptable time for a configuration with many nodes (>100), the simulation uses multithreading with OpenMP. As the RNG functions of OMNeT++ are not thread safe, it is necessary to provide for each thread its own random engine: this is done using the num-rngs option in the INI file. Therefore, it is important to specify the correct number of threads available through this parameter.

Finally, be sure to specify the same number of nodes in the GenTopo.py and the NED file (though the simulation checks this condition).

I did this project while I was an intern at LIP6, so I want to thank all its team.

Special thank to Richard Gardner (@richardg93) who helped me a lot during the beginning thanks to his git but also to Gewu Bu (@GewuBU) who guided me during my internship.