https://github.com/25a0/dcnet-simulator

Simulator to explore and compare various scheduling algorithms for DC-networks
https://github.com/25a0/dcnet-simulator

Last synced: 24 days ago
JSON representation

Simulator to explore and compare various scheduling algorithms for DC-networks

• Host: GitHub
• URL: https://github.com/25a0/dcnet-simulator
• Owner: 25A0
• Created: 2015-10-28T07:15:04.000Z (about 9 years ago)
• Default Branch: master
• Last Pushed: 2015-10-28T07:16:28.000Z (about 9 years ago)
• Last Synced: 2024-10-15T06:08:22.366Z (2 months ago)
• Language: Java
• Homepage:
• Size: 141 KB
• Stars: 2
• Watchers: 4
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
This is the simulator that was used to benchmark the performance of footprint scheduling, 

Pfitzmann's algorithm, and the Herbivore implementation of Chaum's map-reservation algorithm.

Run `make` to compile the simulator. This requires Java 7 (or newer).

After that, run `./simulation.sh` to start the interactive simulator.

The folder `scripts` contains various benchmark scripts that run simulations automatically.

In order to execute a script (e.g. the script `scripts/benchmarks/activity`), enter

	`run scripts/benchmarks/activity`

in the interactive simulator or execute 

	`./simulation.sh scripts/benchmarks/activity`

from the shell.

In order to save the results of a simulation, enter `write` in the interactive simulator once the 

simulation is finished. 

The results will be written to a .csv file in the folder `benchmark-data`. If this file is

already present, the new results will be appended to the existing data.

The plots in the paper can be reproduced with the provided R script `process.R`. This requires

the ggplot2 library. By default, all plots are commented out.

Browsing the predefined scripts gives a good overview of the available options for each simulation.

Below is a quick guideline for running custom simulations:

For footprint scheduling:

	`footprint S B P A C`

	where

		S is the number of slots that can be reserved

		B is the number of bits per footprint

		P is a list containing the number of participants in the simulated networks

		A is the activity rate of all participants as a floating point value between 0 and 1

		C is a boolean that determines whether the discussion should stop once the schedule converged.

			Note that this is only possible in a simulation, not in a real-life scenario where

			footprints are kept secret in general.

	

	So, a simulation of footprint scheduling could look like this:

	`footprint 16 8 [100 200 500 1000 2000 5000 10000] 0.01 false`

For Pfitzmann's algorithm:

	`pfitzmann i P A`

	where

		i is the number of available slots per participant (as discussed in Appendix A)

		P is a list containing the number of participants in the simulated networks

		A is the activity rate of all participants as a floating point value between 0 and 1

	So, a simulation of Pfitzmann's algorithm could look like this:

	`pfitzmann 32 [100 200 500 1000 2000 5000 10000] 0.01`

For the Chaum's map reservations:

	`chaum P A`

	where

		P is a list containing the number of participants in the simulated networks

		A is the activity rate of all participants as a floating point value between 0 and 1

	So, this is how a simulation of Chaum's reservation map could look like:

	`chaum [100 200 500 1000 2000 5000 10000] 0.01`