Skip to first unread message

el9339

unread,
Dec 13, 2017, 11:46:11 AM12/13/17
to ns-3-users
Hi all,

The method to discard packets in NS3 is done through an indication whether the packet is corrupted or not by comparing a random number to the transport block BLER, as it follows (literally copied from Lte-spectrum-phy class):

          (*itTb).second.corrupt = m_random->GetValue()> tbStats.tbler ? false : true;

So my question is the following: Is not it supposed to be the other way around? I mean if m_random is higher than tbStats.tbler value, my understanding is "corrupt" should be true and therefore discard the packet, but please correct me if I am wrong.

Thank you in advance.
Best Regards,

El

Tommaso Pecorella

unread,
Dec 14, 2017, 6:05:11 PM12/14/17
to ns-3-users
Hi,

tbler is, for real, an error rate.
Let's suppose you want an error rate of 10%.
You draw a random number between 0 and 1 and if the random number is GREATER than 0.1, then you do NOT corrupt the packet.

T.

el9339

unread,
Dec 18, 2017, 7:43:00 AM12/18/17
to ns-3-users
Hi Tommaso,

Thank you for answer. Now I see it is a kind of "threshold" and everything makes more sense to me.
However, another question jumps me. Considering PF scheduler algorithm, at the UE side when it is about selecting the CQI to report to the eNB, LSM is used to determined the MCS (and therefore the CQI) according to the sinr to fulfill the bler is less than 0.1. When the UE receives the packet scheduled by the eNB (2 ms after) the bler is 0 or 1 but quite far of its original value at TX. How can it be so degrated? 

BR,
El

Tommaso Pecorella

unread,
Dec 18, 2017, 3:15:47 PM12/18/17
to ns-3-users
The channel noise isn't constant. It is modeled with a statistical distribution, and the node moment can influence it too.

T.
Reply all
Reply to author
Forward
0 new messages