All network simulations involve an element of randomness.
Some examples are -
a. While, it is possible to configure the traffic sources in the simulation to generate traffic in a perfectly regular pattern, this is typically not the case in the real world.
b. Node back-off‟s after collisions are random to resolve contention issues
c. The exact bit which is errored, based on Bit error probability of a wireless channel, is decided randomly
NetSim uses an in-built Linear Congruential Random Number Generator (RNG) to generate the randomness. The RNG uses two seeds values to initialize the RNG.
Having the same set of seed values ensures that the same random numbers are repeated. Thus for a particular network configuration the same output results will be got, irrespective of the PC or the time at which the simulation is run.
Modifying the seed value will lead to the generation of a different set of random numbers and thereby lead to a different sequence of events in NetSim. When simulations are run for a network configuration with different seed values, the results will likely be slightly different.
More advanced users may note that “Confidence” can be established by analyzing a set of results with different seed values for the same network scenario.