[ns] Monitor routing updates

["Chris Develder" <Chris.Develder@intec.rug.ac.be>]

Dear all,

QUESTION:
Is there an easy way to accurately monitor updates of the routing tables (both when using rtProtoDV and LS)?

I guess this will require some modification of the source code, uh? Can I for example modify Agent/rtProto/DV::recv-update to make ns write a line to the tracefile (the one produced by trace-all) if the routing table is modified? (If yes, how do I do this?)

SOME COMMENTS:
This would be useful to try to measure the time needed to recover from network failures: elapsed time between network failure (link going down) and last change in routing tables would be the time needed recover...

(Now, the only thing I can easily measure is the elapsed time between the failure and the last of the routing control messages that is sent right after this failure. (I do this by awk'ing the ns trace.) Clearly, this is just an upper bound for the restoration time, as some of the routing control packets may not affect the contents of the routing table.)

Kind regards,

Chris.