Bit Error Rate

[baBOU16]

Hi,

  As part of my thesis, I had to work with a network simulator to get the Bit Error Rate (BER, a physical layer parameter).

  I worked with NS-2 but I couldn't get the BER with simulation. Now I will try with NS-3.

  Please can anyone tell me if it's possible to get the BER with simulation using NS-3 (or NS-2) and not analytically?

Thanks in advance .

[Tommaso Pecorella]

Hi,

I guess there's some misunderstanding. Both ns2 and ns-3 use a channel and the channel has a BER. What I mean is that the BER is a channel property, and it is a parameter set in the simulation. It may depend on the mutual node's position, the channel model and so on, but ultimately it is something that is calculated using some formulas.

As a consequence, you don't use a network simulator (one like ns-3 or ns2) to find the BER, because the BER is known.

Hope this helps,

T.

[Basma Bouraoui]

Thanks for your response.

I have another question please about Packet Error Rate (PER).

Could you please tell me if it is possible to get the PER according to the BER (set in the simulation) from the trace file? If isn't possible could you recommend me another network simulator (OMNET, OPNET, …) to get the BER or the PER with simulation?

[Tommaso Pecorella]

Hi,

again, PER is a function of BER. It may be as simple as a binomial function or much more complex (if there is a FEC). Still, it's algorithmic.

The usual procedure is:

1) you know the nodes positions

2) you know the channel model

3) you know the antenna model, the antenna characteristics, the Tx power, the Rx noise figure and so on.

4) you calculate the signal and the noise power at the receiver (there may be a random factor to consider the model randomness).

5) you calculate the BER

6) from the packet's length and FEC, you have a probability that the packet is correctly decoded.

7) you throw a random number and you decide if the packet is correctly received or not.

As a consequence, you can use a network simulator to verify that the PER is the expected one. Or, if the number of factors to consider are too high (i.e., the PER considering the packet collisions) you can find the experimental PER.

A common case is in CDMA, where the PER is dependent on the number of concurrent transmission (non-perfectly orthogonal codes, ISI) and the higher the PER, the higher the network load, the higher the ISI (positive feedback).

T.

[Sejal Haveliwala]

How to integrate the Loss differentiation algorithm with any TCP protocol like TCPnewreno,tcpwestwood and others,I want to compare the accuracy and throughput ..Please someone help me .

[Tommaso Pecorella]

Please read the posting guidelines.

T.