Tools for data analysis on Omnet++

The following tools are provided:

• install_tools.py
• update_template.py
• make_runfile.py
• parse_data.py
• analyze_data.py
• scan_datafile.py

install_tools

Script used to install the present tools in the Omnet++ directories. The Omnet++ environment variables must be set (invoke .setenv script in Omnet++ before attempting to install). The tools will be installed in the bin/ directory of the Omnet installation in the form of symlinks.

update_template

Tool to create files from a directory containing .mako templates. All found .mako templates will be processed. The name of the created file will be based on the template name, stripping the .mako extension. For example:

• file1.ned.mako -> file1.ned
• file2.ini.mako -> file2.ini

If the file already exists, the file will be re-created only if the template is newer than the file or if the -f switch is used.

Relevant parameters:

• -h, --help: show help and exit
• -f, --force: force update
• -r, --recursive: enable recursive search of tempalted in sub-directories
• -d DIR, --dir DIR: directory to work on (default ./)
• -c CONFIG, --config CONFIG: YAML Configuration file (default None). The dta structure of the YAML file will be made available in the configuration dictionary to all the templates processed. Useful if multiple templates must rely on the same configurations (e.g., .ini files and .json file for data analysis)

make_runfile

Create Runfile (it is basically a Makefile) to tun multiple Omnet++ experiments in parallel (using the make command and its -j option). The script reads the configuration for a Omnet++ .ini file and generates the command to run all the experiments for every considered configuration (excluding configuration containing the NORUN string in their name).

Relevant parameters:

• -h, --help: show help and exit
• -z, --zip: The output of the simulation (.sca file) will be automatically compressed with gzip. This parameter can save lot of disk space when running large experimental campaigns
• -o OUTPUT, --output OUTPUT output file (default: Runfile)
• -f FILE, --file FILE input file (default omnetpp.ini)

parse_data

Parse a set of .sca files (related to multiple runs and configurations of the same experiment) and store the relevant results in a database.

The results are organized as scenarios and runs of the same scenario. A scenario is identified by a set of parameters that set the scenarios apart from each other. For example if we are studying a server load curve the scenarios can be set apart by the rho load parameter or by the lambda incoming request rate. For each experiment we collect several metrics that can be aggregated (for example if we have several computing nodes, the utilization metric would be referred to each node, but we may be interested in collecting only the average a standard deviation of this metric).

The tools can also collect data in the form of histograms

Relevant parameters:

• -h, --help: show help and exit
• -r, --reset: recreate database
• -c CONFIG, --config CONFIG: .json config file with the definition of every scenario and metric (default config.json)
• -d DB, --db DB: database file (default test.db)
• -j JOBS, --jobs JOBS: number of parallel tasks to parse .sca files (default 1)
• F: list of .sca files to process

analyze_data

Second step of data processing tools. From the database crated with parse_data we create a set of data files ready to plot or to further process with pandas or R. The configuration is in the same .json file used for parse_data. For each analysis we describe the output files, the parameters in the scenario that are to be kept fixed and the scenario parameter to scan (that will be used to crate the rows of the data file). The output will have a set of column describing the trajectory in the scenario space and a set of columns for the selected metrics. For each metric we provide the value (averaged over the runs of the same scenario) and the standard deviation of the parameter over the runs (sigma-labelled column).

Also histograms can be extracted from the database.

Relevant parameters:

• -h, --help: show help and exit
• -c CONFIG, --config CONFIG: .json config file with the definition of scenarios, metrics and analyses (default config.json)
• -d DB, --db DB: database file (default test.db)
• -v: verbose mode

scan_datafile

Simple tools to list the columns of a data file created by analyze_data.