Hi Joahannes,

Thank you for the clarification.

Actually our idea of the problem was very similar to what you said: from the logs we could see that the TCP connection would fail and there would be no new attempt at sending the bundle again later (there was no new checking of the routing tasks to be done for the done even thought beacons were being received). We would not have thought that this was related with DHT, we simply thought the code was not ready to deal with this kind of situations (i.e., TCP fails but the node is kept as neighbor), but we were suspicious since another colleague had succeeded with IBR-DTN in similar situations (using a previous version).

One question, as we are not familiar with DHT: what should be the consequences of disabling DHT? For now in a small lab environment it all seems to work well without DHT, but should we expect problems once we move into a real-world environment with hundreds of nodes?

Thank you,

Romeu