Analysis Scripts for OMNeT++
These analysis scripts are for OMNeT++ 4.0 vectors and scalar files. They are based on the scavetool (an OMNeT++ command line utility to process vectors and scalars), bash commands (awk, tail, cat, sort, dos2unix, etc), gnuplot, bargraph and the R-project (for compute some statistics). THEY ONLY WORK ON LINUX, and THEY DEPEND ON ALL THE PROGRAMS I MENTIONED. If you miss one, the scripts will not work.
All these scripts ASSUME the existence of a bash environment variable called OMNET_PATH indicating the path where the OMNeT++ is installed. I.e: if omnet is installed in the /opt/omnet40 directory, the OMNET_PATH (in bash) should be :
$ export OMNET_PATH=/opt/omnet40
Leyend
- Vector: serie of data collected over the time.
- Scalar : single values collected along the simulation.
- Statistics: descriptive information about an scalar. It considers mean, min, max and std.dev.
- Replica : single simulation execution. The only difference among replicas within the same experiment is the Random Number Generation seed.
- Experiment: set of replicas (vector and scalars files) coming from the same configuration in the simulator, and ONLY varying the RNG seed.
General Scripts Usage
The scripts are separated in three groups: first, to extract data from simulation output files; second, to integrate that extracted information into a single file; and analysis scripts, to perform analysis on the data.
The first group of scripts will generate different types of outputs (directories, temporary files, or data files). Integration scripts take temporary files and integrate them into a single file. We call this these data files. All analysis scripts are based on the data files format. Thus, the procedure to analyze some output is the following:
- extract the data you need (vector, scalar or statistic)
- integrate them into a data file.
- execute the analysis to the data file.
For all the examples shown, we assume an experiment called "Example" (it comes from the config name in the omnetpp.ini file). We have several replications on this experiment. So, we have 3 files on each replica. I.e. for replica 0, we have Example-0.vec, Example-0.vci, Example-0.sca (vectors time series, vectors resume, scalars and statistics). So, the experiment name is Example (the base name of each file, removing the hyphen and the replica id.
Extraction Scripts
Data Integration Scripts
Analysis Scripts