Hello.

Am 11.11.2013 16:55, schrieb Luís Carlos Guedes:
Hello, 
Can you please tell me where can I see(in the code) the place where you did that configuration; 
"If A meets B again(for the 2nd time) the remaining fragment X' is forwarded. In case node A meets another peer, the complete bundle X is forwarded."

This is the default if you enable fragmentation ("fragmentation = yes").

Since reactive fragmentation only works with convergence-layers supporting that feature, you can find the code to do that within the TCPConnection.cpp.

https://github.com/ibrdtn/ibrdtn/blob/master/ibrdtn/daemon/src/net/TCPConnection.cpp#L655

Because in order to send information from A to B through 2 different nodes (and join them in the node B) i need to send through the second node just that X' and not the complete bundle as is already implemented.

You can only do this if you know, that the fragment X'' do not get lost on its path and will definitely reach its destination. In most DTN scenarios you would not know that.

Besides that i wanted to know how do you implement in the code to join the information (from 2 or 3 fragmentations), to they become just one bundle like in the beginning.

The bundle merge is done in the FragmentManager. This process runs asynchronously to not slow down the forwarding process.
https://github.com/ibrdtn/ibrdtn/blob/master/ibrdtn/daemon/src/core/FragmentManager.cpp#L81


Kind regards,
Johannes Morgenroth