Sviluppo di una tecnica di riconoscimento statistico di applicazioni su rete IP

These are the classes that implement the studied method of identification exploiting all the aforementioned auxiliary classes. They are Learn-Rec.java, LearnRec2.java, LearnRec2_mod1.java, LearnRec2_mod2.java and LearnRec2_mod3.java. The latest three ones are only an attempt to save time subdividing LearnRec2 into the three parts corresponding to the three probabilistic modes (described in chapter 4); so they won't be taken into account any more, being their code only a copy of parts of LearnRec2.

LearnRec.java is the class that realizes the identification process calculating the pdf values only for the unknown PktVariables/SessionVariables passed to getFValue(). The first part, the learning part, of its main method can be subdivided into these subsections:

• parameters’ acquisition and initialization of counters for percentage calculation;
• creation of a trace object with all the sessions from the '.txt' learning files;
• selection of the sessions with at least dim_min packets and duration lesser than DUR_MAX. This leads to the creation of a reduced Trace object;
• determination of the number of size and time bins and selection of the binning grid (linear or logarithmic) ;
• creation of '.txt' files with information contained in the reduced Trace for debug's purpose;
.creation of a Trace object containing sessions to identify;
• selection in this Trace object of the sessions with a number of packets larger than dim_min ..