DTN and AX.25

Almost 12 months ago Darren, G0HWW sent this post to the linux-hams mailing list. We exchanged some emails discussing how Delay Tolerant Networks (DTNs) could be used for Emergency Communications (being a classic store and forward network). In a DTN network, each sub-net or point-to-point link can operate over whatever stack is available, in this case it is AX25. In August Darren became aware of some work done in Helsinki University of Technology (under Joerg Ott) specifically a paper entitled Opportunistic Email Distribution and Access in Challenged Heterogeneous Environments. Joerg forwarded Darren their DTN Mail Proxy and we started experimenting with it.

Last weekend, we went a step further. Darren sent me on patches to the DTN2 reference implementation that implement an AX25 Convergence Layer. As I had a working 9600 baud AX25 connection we were eager to test it. After some patching, head scratching and recompiling we were able to successfully pass email from my Laptop over AX25 (439.850MHz) to my Linux server, then over the Internet to Darren’s server and on to his mailbox. All using the DTN bundling protocols. Very cool!

TM-D710E and KISS mode.

Happy New Year!

Now Christmas is over, back to business. I spent some time last Saturday testing KISS mode on the TM-D710 to see if Kenwood have improved it. The test scenario was.

Linux Laptop <-> TM-D710E  <-> 439.875MHz <->  SCS-PTCIIex <-> FT-847 <->  Linux Server.

At both sides, the serial ports were running at 57,600bps. I configured both sides to use 9600bps over the RF connection, I used settings of txdelay 100ms, Paclen of 255, maxframe of 7, mtu of 1500, mss of 1460, and tcp window of 2920 as suggested here. Other settings were at their defaults.

I started an FTP download of a (roughly) 1.6MB file from the Linux Laptop to the Linux server, and about 74 minutes later, it completed.  I then downloaded the same file and approximately 70 minutes later it finished.  A quick comparison of the md5 checksums showed the file to be ok.  During the transfers, it seemed to me that the 710 was a bit ‘slow’ in switching from receive to transmit (compared to the PTCIIex which was almost instantaneous).

That said, given my very limited experience with 9600 baud packet, I still think it is respectable, though more testing is needed and I would welcome any suggestions for different parameters to change.