p2pandaHui! This is #p2panda doing #localfirst sync over #LoRa in a #mesh network 🐼 📻
We're experimenting currently with a log-based, replicated delay tolerant / store-and-forward protocol using the RNode transceiver firmware. All very early but fun!
That Vasket@p2panda tow of my favorite #p2p things working together. That's really neat! Do you only use an Rnode or also using #Reticulum ?
@thatv Right now we're "just" using the RNode (it's really well done!) with our own replicated routing protocol on top.
We still need to evaluate if using p2panda-encryption is a good candidate to build a broadcast "transport encryption" for meshes, so you can do multicast for managed groups with FS and PCS over LoRa. Also our initial unicast key agreement is async as we use X3DH while Reticulum assumes that peers need to be "online" for secure links.
@thatv It's too early to say for now, we have a lot of reading to do but we really love Reticulum and it's quality.
We have the feeling that maybe, even if we wouldn't use the full power of Reticulum's cryptographic features (as we would bring our own, "more delay tolerant" and "more broadcast" ones) it should be possible to run this on top of it in any case (even if it's just using the "plain" destination type).
@p2panda this is a good approach I think and I think "more delay tolerance" is always a good idea. As far as I understand Reticulum is more of a better (encrypted/privat) TCP/IP stack with organizes it self via UDP like packes. Read the docu (around 0.8.x iirc) some time ago so maybe things changed. If you want a deylaytolerant store and forward mechanic on top the Reticulum. You could use the #lxmf propagation nodes. But that's just an idea.