To run the experiment, make sure the following files are under the same directory:

• experiment_setup.py
• experiment_setup.sql
• allIGP
• AS_links
• ingress_egress
• peering_links
• rib100.txt, rib10000.txt, rib1000000.txt

Rib data obtained from Routeviews:

http://archive.routeviews.org/bgpdata/2003.01/UPDATES/updates.20030101.0005.bz2

It was preprocessed using the following bgpreader command: bgpreader -i -d singlefile -o rib-file,http://archive.routeviews.org/bgpdata/2003.01/UPDATES/updates.20030101.0005.bz2 | head -n <number of lines>

./experiment_setup.py

• MIRO experiment

  Running the experiment using the MIRO method:

  exp = Experiment()                                # Initialize experiment and database
  exp.setup(1239, 2917)                             # Create tables for given downstream, upstream AS
  exp.load_strawman()
  
  exp.add_policy('216.79.72.0/24', 6389)            # Add policy to avoid prefix and AS path
  exp.load_bgp('rib100.txt', 'insertiondelay.log')  # Load RIB file, time insertions, and output to file

• Wiser experiment

  Running the experiment using the Wiser strawman method:

  exp = Experiment()                                # Initialize experiment and database
  exp.setup(1239, 2917)                             # Create tables for given downstream, upstream AS
  exp.load_wiser()                                  # Load Wiser policy checking triggers
  exp.load_bgp('rib100.txt', 'insertiondelay.log')  # Load RIB file, time insertions, and output to file
  exp.unload_wiser()                                # Unload the policy checking triggers when experiment finished

Running the experiment using the residue method:

exp = Experiment()
exp.setup(1239, 2917)
exp.load_residue()

exp.add_policy('216.79.72.0/24', 6389)
exp.load_bgp('rib100.txt', 'insertiondelay.log')

The experiment_RAM_usage.py measure the memory usage using the top command. the run function in the experiment_RAM_usage.py is the function that runs the experiment and collects the memory usage data. Given an experiment setup (the downstream AS, the upstreamAS, scenario and the BGP feeds number) and the number of points to collect, say N, the run function will issue the top command after every (BGP_feeds_number\N) BGP feeds consumptions, and write the result to the <scenario>_ram.log file in the Gnuplot\log folder.

For example, the following codes run the MIRO experiment with the residue method using 1000,000 BGP feeds, and then collects the memory usage for every 10,000 BGP feeds consumptions.

downstreamAS = 2914
upstreamAS = 1239
experiments = {
    'miro_residue':('residue','miro',20,1000000),
    }
for exp, data in experiments.iteritems():
    run(upstreamAS, downstreamAS, exp, data, 100)

• BGP feeds processing delays

  Filename format: <downstream AS>_<upstream AS>_<operation measured>_<BGP feeds read>.log

  Data format: <BGP feed row number> <Processing delay in milliseconds>

• Memory usage

  Filename format: <scenario>_ram.log

  Data format: Every two lines form a data point. The first line gives the number of BGP feeds processed, and the second line gives the information returned by the top command, which has the following fields:

  PID, USER, PR, NI, VIRT, RES, SHR, S, %CPU, %MEM, TIME+ COMMAND

  We are interested in the RES filed which stands for the physical memory taken by current connection to the database. For the meaning of other fields, please refer to the top command manual.