DYMO simulation - DYMONetwork.net / A

[Louis Boile]

    Hi everybody,

I am trying to simulate a mobile ad-hoc network using the DYMO routing protocol. I'm using Omnet++ 4.4.1 and Inet 2.4.0.

I am simulating a simple scenario with 6 nodes and I am using BonnMotionMobility in order to specify their mobility - with the attached UAV.text file.
Host[0] is fixed and wants to communicate with a moving Host[1] during all the time of the simulation. Host[2], [3], [4] and [5] are here to maintain the communication link between them. I attached a gif file showing the mobility of the nodes.

I was using first the network using DYMONetwork.ned given in inet/examples/manetrouting/dymo. I attached the omnet.ini file (omnetpp_DYMONetwork.ini) I am using. The radio parameters are set up for a 20km range with a Free Space Propagation Model.

At the beginning of the simulation, there is no intermediate host to maintain communication between hosts[0] and [1] and no ping replies are received. Then between t=400s up to 1800s DYMO routing specifies a connection using host[2] as intermediate node.
The simulation is working up to t=1800s. At this time nodes don't receive any ping replies. I don't think any RERR message is sent whereas one intermediate host is no longer sufficient to maintain communication between hosts [1] and [0]. Do you know what is happening there? I think there is a problem with the RERR messages and in the statistics I don't see any of the messages sent by DYMO but I am not sure. I attached screenshots of the round-trip-time vectors of hosts [0] and [1] (rtt_host[*]_DYMONetwork.png).

I have seen another network Net80211_aodv.ned in inet/examples/manetrouting/net80211_aodv. I am using the same settings as before with DYMO routing protocol The omnetpp.ini file attached :omnetppNet80211_aodv.ini. I am having the following results with ping app:

    Net80211_aodv.host[0].pingApp[0]
--------------------------------------------------------
ping host[1](ipv4) (145.236.0.2):
sent: 5000   received: 30   loss rate (%): 99.4
round-trip min/avg/max (ms): 139.302/1335.75/3040.03
stddev (ms): 738.077   variance:0.544757
--------------------------------------------------------
--------------------------------------------------------
    Net80211_aodv.host[1].pingApp[0]
--------------------------------------------------------
ping host[0](ipv4) (145.236.0.1):
sent: 5000   received: 32   loss rate (%): 99.36
round-trip min/avg/max (ms): 1005.16/1342.54/3012.19
stddev (ms): 587.378   variance:0.345012
--------------------------------------------------------

I attached the round-trip-time of host[0] and [1] (rtt_host[*]_net80211_aodv.png). This time DYMO seems to update the routes but the rout-trip-time is very high - 1300 ms in average and almost all the ping messages are lost.

Does anyone knows what is incorrect in the simulations? Or does anyone knows why there are two ned models for DYMO?

Thank you very much