Gossiping with Append-Only Logs in Secure-Scuttlebutt

This project implements the gossiping with append-only logs in secure-scuttlebutt, as specified in the assigned paper

Requirements

• Java 11
• Eclipse IDE 2020-06 (4.16.0)
• Repast Symphony 2.8.0

Optionally, for running the plotting scripts:

• Python 3

Installation

You have two ways to do it; first one:

1. Download the installer.jar file from this link
2. Execute it (java -jar installer.jar) and follow the installation wizard
3. Run the start_model.bat or start_model.command file

Second one:

1. Clone the repository
2. Import the project as Repast Project into Eclipse
3. Run the simulator

Simulator parameters

• Num LANs: the number of LANs that will be present on the simulation. This is a fixed parameter that cannot vary during the run of the simulation.
• Num People: the number of Person actors in the simulation. Similarly to the previous parameter, this number cannot change during the simulation run.
• Motion Strategy: the different motion strategies are described in paragraph \ref{subsectionmotion} and define the way the Person actors move in the simulation's grid.
• Synchronization Protocol: the synchronization protocol defines how events are exchange between persons. The two different ways of synchronizing event are explain in detail in paragraph \ref{subsectionsync}.
• Gaussian mean preferred LANs: used in the Habit Motion strategy, indicates the mean number of favourite LANs that a persons has.
• Gaussian std preferred LANs: used in the Habit Motion strategy, indicates the standard deviation number of favourite LANs that a persons has.
• Gaussian mean ticks waiting: used in the Habit Motion strategy, indicates the mean number of ticks a person will stay in one of his favourite LANs (and at home).
• Gaussian std ticks waiting: used in the Habit Motion strategy, indicates the standard deviation number of ticks a person will stay in one of his favourite LANs (and at home).

Plotting scripts parameters

countEvent.py

Simple script that plots the number of events created during the simulation.

• --file: input log file
• --group: rate at which ticks are grouped (useful for having a cleaner graph with longer simulations)

countEventDelivery.py

Simple script that plots the number of events delivered during the simulation.

• --file: input log file
• --group: rate at which ticks are grouped (useful for having a cleaner graph with longer simulations)

diffusion.py

Script that plots the number of time each event has been delivered during the simulation.

• --file: input log file

latency.py

Script that plots the latency of the different events during the simulation.

• --file: input log file
• --lans: number of LANs in the simulation
• --persons: number of persons in the simulation
• --mu_lans: mean of the normal distribution of the preferred LANs
• --std_lans: std of the normal distribution of the preferred LANs
• --mu_wait: mean of the normal distribution of the waiting time
• --std_wait: std of the normal distribution of the waiting time
• --target: csv to append result to. If this parameter is specified then the graph won't be plotted.

plot_latency.py

Script that plots the 3d graph of LAN/Persons/Latency.

• --file: input log file
• --type: type of plot in [Basic, MeanPrefLANs, MeanWait]

Authors

• Simone Degiacomi 211458
• Davide Tessarolo 211457
• Davide Zanella 211463

Reference

Anne-Marie Kermarrec, Erick Lavoie, and Christian Tschudin. 2020.Gossiping with Append-Only Logs in Secure-Scuttlebutt. In1stInternational Workshop on Distributed Infrastructure for CommonGood (DICG’20), December 7–11, 2020, Delft, Netherlands.ACM, NewYork, NY, USA, 6 pages.