README

We use python as our DSL for specifying experiment configuration

The module must define one attribute:
  control_flow => a sts.control_flow.*.ControlModule object.

This class takes a SimulationConfig objects as its first parameter.

SimulationConfigs can optionally take the following contructor parameters:
  topology_class    => a sts.topology.Topology class (not object!)
                       defining the switches and links
  topology_params   => Comma-delimited list of arguments to pass into the FatTree
                       constructor, specified just as you would type them within
                       the parens.
  patch_panel_class => a sts.topology.PatchPanel class (not object!)
  dataplane_trace   => a path to a dataplane trace file
                       (e.g. dataplane_traces/ping_pong_same_subnet.trace)
  violation_persistence_threshold => number of logical time units to observe a
                                     violation before we declare that it is
                                     persistent
  switch_init_sleep_seconds => number of seconds to wait for switches to
                               connect to controllers before starting the
                               simulation. Defaults to False (no wait).
  monkey_patch_select => whether to use STS's custom deterministic
                         select. Requires that the controller is
                         monkey-patched too

Config files may also optionally define a function `get_additional_metadata`,
that takes no arguments and returns a dictionary storing metadata about the
experiment (e.g. the current commit hash of the controller under test).