Bug 2825 - vanet-routing-compare can not track Tx packets (and more).
vanet-routing-compare can not track Tx packets (and more).
Product: ns-3
Classification: Unclassified
Component: wave module
All All
: P3 major
Assigned To: ns-bugs
Depends on:
  Show dependency treegraph
Reported: 2017-11-18 19:48 EST by Tommaso Pecorella
Modified: 2017-11-18 20:32 EST (History)
0 users

See Also:

preliminary patch (5.29 KB, patch)
2017-11-18 20:32 EST, Tommaso Pecorella
Details | Diff

Note You need to log in before you can comment on or make changes to this bug.
Description Tommaso Pecorella 2017-11-18 19:48:10 EST
In the example there are two places where the code is not congruent.

VanetRoutingExperiment::ConfigureDefaults ()
  Config::SetDefault ("ns3::OnOffApplication::PacketSize",StringValue ("64"));
  Config::SetDefault ("ns3::OnOffApplication::DataRate",  StringValue (m_rate));

... but there's no OnOffApplication in the code. Instead, BsmApplication is used.

VanetRoutingExperiment::ConfigureApplications ()
  oss << "/NodeList/*/ApplicationList/*/$ns3::OnOffApplication/Tx";
  Config::Connect (oss.str (), MakeCallback (&RoutingHelper::OnOffTrace, m_routingHelper));

Again, no OnOffApplication in the code.

The second point requires more changes in the code because BsmApplication doesn't have any "Tx" trace.

This has been spotted in this thread:
Comment 1 Tommaso Pecorella 2017-11-18 20:32:23 EST
Created attachment 2961 [details]
preliminary patch

The patch still doesn't work, but there are more issues.

The MAC/PHY overhead is totally wrong. In the original code it gave numbers around 3000 (impossible for a ratio).
It was an unsigned int subtraction round problem. Fixing it gives a negative number, and this means that the calc is wrong to the boots. Practically it's getting numbers from questionable sources.

Some real work is needed (and tbh I don't have much spare time).

The current issues are:
1) I had to remove the "rate" parameter from the example. BsmApplication is fixed-rate.
2) The Trace I added isn't hooked properly.
3) The statistics are really messed up. We need to double check them.