Hi,
the application (i.e., the traffic model) is one of your choices. It will affect the PLR, probably, but there is not one "better" than the other.
About the other question, I think you have a few confused ideas.
A packet can be dropped for one of these reasons (for the 1000th time):
- Channel errors
- Queueing reasons
- Routing decisions
First case: the MAC is dropping it because there are too many erroneous bits. The MAC Drop Trace is fired.
- OR - the Mac is giving up transmission due to channel congestion. The MAC Drop trace should be fired in the Tx node.
Second case: a queue is full. The Queue Drop Trace is fired.
Third case: the routing finds out that there's no route to destination. The IP Drop trace is fired.
Each d**n trace:
- is in a different class,
- the name (usually) is "Drop", but it's not mandatory
- what they're named and what they are depends on YOUR simulation setup
Last but not least, "Drop" can be measured:- link-level (between each pair of nodes)
- end-to-end (from source to destination)
Now, clarify yourself what are you willing to measure, what is your experiment and so on. Then you should hook all the relevant traces and measure the "drops".HOWEVER, to have a "drop rate", you'll also need the number of Tx attempts...
T.