• Prerequisites

  • python3 (I used 3.10.2)
  • networkx install on python3
  • docker

• To run a simulation and generate a dataset, ust run: python3 master.py <number of nodes> [nsfnet/geant2/germany50] <number of runs per traffic intensity value> <time per run in SECONDS> <frequency of statistics in MILLISECONDS> <min traffic intensity> <max traffic intensity>

  • the second argment: nsfnet/geant2/germany50 is the type of network, each one more densely connected than the other
  • the master.py orchestrates all the terminals to run the simulation
  • if:
    • number of runs per traffic intensity value is x
    • min traffic intensity is i
    • max traffic intensity is j
  • then:
    • the script runs a simulation x times for every value in the range [i, j] (both inclusive)
    • each run has the same test, same target bandwidths between runs
    • between different traffic intensities, the target bandwidths of connections vary
    • for every run, the topology and the link speeds of the links stay the same

• Here is the overview of what master.py does:

  • for each simulation (uasge of master.py), we start by randomly generating a topology and the required mininet topology file & ONOS config files

    • this is what the file main.py does, it is used once per simulation
    • this means that the topology and the bandwidths of the links in the virtual network are same throughout the entire simulation
    • we have a range of traffic intensity/lambda values that are also arguments to master.py (we collect a min and max traffic intensity value)
    • for each traffic intensity/lambda value in the user-provided range:
      • we generate a test file that uses d-itg commands to generate traffic between each pair of nodes, and record results
      • we do this once for each traffic intensity/lambda value, since its value determines indirectly target bandwidths for each pair of nodes in the test
      • then, we perform a series of runs for a given traffic intensity/lambda value; the number of runs is again an argment for master.py
      • we prepare for a series of runs by first setting up ONOS
        • this is only done once for a set of runs (for a given traffic intensity/lambda value)
        • ONOS is run in a docker container, to keep things as portable as possible
      • then, we set up mininet
        • mininet is also run in a docker container, to keep things as portable as possible
        • the mininet container also has d-itg installed for the test script
        • note that we setup ONOS and mininet in tmux windows to control them programatically
      • then, we run the test script n times, n being specified by the user
        • between each run, we use sed to change where the output of the test run is sent to
      • just before the last run has executed ( seconds before it finishes), we query ONOS to get the path between each pair of hosts.
        • the script for this is generated by main.py
      • and then we tear down the ONOS & mininet containers

  • the files are generated by main.py and main_extension.py

    • main.py generates everything except the test script test.py

  • look at the function run_setup in master.py to get the idea of what happens for a round of a simulation

  • the function runner does set up for a simulation

  • please note that the file "test.py" changes between runs: specifically the locations where the output raw filed and decoded files change

  • the output should look like:

   simulations/
   	<network_type>_<# nodes>_<min traffic intensity>_<max traffic intensity>/
   		custom_topo.py
   		get_hops.sh
   		graph_attr.txt
   		onos_config.json
   		sorted_links.txt
   		topo.json
   		lambda_9/
   			run_1/
   				raw_data/
   					2024-11-29-19:10:46.121925_4_4_0_1 
   					2024-11-29-19:10:46.121925_4_4_0_2 
   					2024-11-29-19:10:46.121925_4_4_0_3
   					...
   				decoded_data/
   					2024-11-29-19:10:46.121925_4_4_0_1 
   					2024-11-29-19:10:46.121925_4_4_0_2 
   					2024-11-29-19:10:46.121925_4_4_0_3
   					...
   			run_2/
   				...
   		lambda_10/
   			run_1/
   				raw_data/
   					2024-11-29-19:15:24.290317_4_4_0_1 
   					2024-11-29-19:15:24.290317_4_4_0_2 
   					2024-11-29-19:15:24.290317_4_4_0_3
   					...
   				decoded_data/
   					2024-11-29-19:15:24.290317_4_4_0_1 
   					2024-11-29-19:15:24.290317_4_4_0_2 
   					2024-11-29-19:15:24.290317_4_4_0_3
   					...
   			run_2/
   				...