It may not come as a surprise that the network works after it was working just 
fine before. The big deal here is that we are now using the fancy Ethernet 
switch to create multiple VLANs to isolate critical data paths.

With multiple networks, we need something to route data between them. Thanks to 
the magic of tagged VLANs (802.1Q), all of these VLANs are delivered over one 
physical interface to the Intel NUC. So now the NUC is functioning as the core 
router on top of telemetry and other duties.

With this change we were able to switch to a dedicated link between ground and 
launch tower. Hopefully that will help with some of the tower communication 
issues we have seen in the past. A dedicated VLAN has been created for 
non-critical ground WiFi so that service can easily be disabled if 
communication problems pop up.

Most importantly with all of these different VLANs, we will be able to quickly 
narrow down where network issues are occurring. This should put us in a good 
place to spend more time squashing bugs than hunting for them.

Immediate concerns:
- Most off-rocket hosts have new IP addressees.  Any associated hard-coded 
addresses will need to be updated.
- Live AIS data is not displaying at the FDT. Maybe a socat command on the tm3k 
needs a fresh IP?

Future work:
- setup Nagios
- setup remote syslog everywhere
- install Splunk
- win!

On May 22, 2015 6:31:26 PM PDT, Andrew Greenberg <> wrote:
>Kenny got the entire FDT/LTC/FC network stack up and running. After
>stupid configuration problems with the ubiquiti POE-to-wifi adapter, it
>now just all works. We even power cycled it and made sure it came up
>sanely. Go Kenny!


psas-avionics mailing list

Reply via email to