I can not test this at the moment since I have no 64 bit version of linux running at this time. Also as Cereal said, a graphical environment is needed to run this. As the most stable linux servers don’t run in an graphical mode, this is not very interresting (by the way, it’s never interresting to have a server running in a graphical mode since this takes a lot of processing power away only to see some visuals + could make the system more instable).
Also an other notice, some of the (embedded) boards I’m working on even don’t have a video out, or aren’t meant to use with a display (for ex. the olimex olinuxino imx233 or the a13 micro).
If you want to have your server app running stable, you’ll have to split it up in 2 parts. The “real” server part, which takes care of the connections an distribution of the velbus packets and is a console app that runs in the background. The visualisation part, which would be optional and has only no more to do as connecting to the server part and to do some visuals. This as reason that when the visual part crashes the whole system doesn’t goes down and it will be usable to users that don’t use a display.
Why is it so interresting to use embedded boards with no display: simply because they are so low power consuming, very stable and there are boards that will fit in a din rail enclosure (look for ex at tuxrail from busware.de). For domogik I’d recommend using a rasperry pi at this time since the other boards don’t have enough memory (waiting for other boards to try out, but the last few months I didn’t had the time at all).
So, I hope this is a bit more clear why your app isn’t that interresting on linux at this time, yet it looks very promessing on windows.
ps: another advantage when splitting up your app, you could have the server part running at a linux station (for ex.) and have the front app (the monitoring part) still running in windows (I’m not doing everything in linux btw ;-))