**3.1. Updates synchronization—classical experiment**

Floyd Sally and Van Jacobson have made an experiment [17] in which they proved that without a randomization of the duration of the update counter, all the devices tended to synchronize their update periods, thus periodically overloading the network.

With the same goal, we can re-validate their experiment much faster, by using the HypeR‐ SimRIP application. Even though the concept of the application requires by default the randomization of the 30 sec update period with a random value between −5 and 5, the application allows the user to modify these limits.

In the HypeRSimRIP application, in order to deactivate the above-mentioned randomization, the user must set equal values for minimum and maximum randomization limits (preferably bothvalues shouldbezero)intheSettingswindow(seeFigure10)beforestartingthesimulation.

For the experiment, consider 10 routers, in the configuration given by Figure 11, with a full graph generated between them (connections between every router), generated by using the 'Full Graph Generation' options from the 'Diverse' menu of the application (see Figure 12).

After assigning the IPv4 addresses, all the routers are activated via the 'Set Online/Offline' window (Figure 7).

Commencing the simulation, devices start transmitting RIP datagrams and the application will start to show the network usage (see Figure 13). Periodically, huge network usage explosions


**Figure 10.** Settings window—randomization deactivation.

**Figure 11.** Ten router topology.

are observed: at first 484.496 bps, then periodically (every 30 sec) 171.360 bps. By using the temporal acceleration facility, it can be observed that this tendency will remain throughout the simulation of this experiment.

By repeating the experiment, with the randomization margin between −5 and 5, the network usage from Figure 14 was obtained.

#### A Virtual Routing Solution for IP Networks http://dx.doi.org/10.5772/65415 71

**Figure 12.** Topology determined by a full graph between 10 routers.

**Figure 13.** RIP network usage without randomization (first minute).

**Figure 14.** RIP network usage with randomization (second minute).

are observed: at first 484.496 bps, then periodically (every 30 sec) 171.360 bps. By using the temporal acceleration facility, it can be observed that this tendency will remain throughout the

Proceedings of the International Conference on Interdisciplinary Studies (ICIS 2016) - Interdisciplinarity and Creativity

By repeating the experiment, with the randomization margin between −5 and 5, the network

simulation of this experiment.

**Figure 11.** Ten router topology.

usage from Figure 14 was obtained.

**Figure 10.** Settings window—randomization deactivation.

in the Knowledge Society

70

**Figure 15.** IP range window.

Most of the values during the update periods remained at 17.136 bps (10 times less than the minimum usage boom from the without randomization simulation), and in just a few excep‐ tions, the network usage reached 51.408 bps (still half the minimum usage boom from the previous simulation).

In conclusion, because the transmissions are distributed, the network is not anymore disrupted significantly by the RIP datagrams, thus avoiding a possible network overload.
