lmfeeney
commented
9 years ago In the MAC and application Castalia code, we should turn off most of the traces.
In general, I think the easiest thing to do (feel free to disagree!) would be to enhance the Castalia executable with actual OMNeT++ statistics collection. Then we simply ignore the Castalia output, which isn't really serving our needs well -- especially compared with what modern omnet gives us.
I don't think we need too much data, at least in the first instance.
- a vector of application packet sequence numbers (indexed by time)
- a vector of association and dissociation times
- a vector of beacon arrivals from own coordinator (indexed by time)
- a vector of buffer sizes (collected each time a pkt is added or removed)
I think that file format will be easier to deal with...omnet -> scave -> gnuplot/whatever.
Laura
On 12/15/14 14:00, Michael Frey wrote:
We should clean up most of the code since there is a lot of redundancy. Also, the usage should be eased - the command line options are actually a mess.
— Reply to this email directly or view it on GitHub https://github.com/mfrey/cthulhu/issues/1.
We should clean up most of the code since there is a lot of redundancy. Also, the usage should be eased - the command line options are actually a mess.