We propose an original approach which allows the characterization of network communication graphs with the network motifs. As an example we checked our approach by the use of network topology analysis methods applied to communication graphs. We have tested our approach on a simulated attacks inside a scale-free network and data gathered in real networks, showing that the motif distribution reflects the changes in communication pattern and may be used for the detection of ongoing attacks. We have also noticed that the communication graphs of the real networks show a distinctive motif profile.