alt Hacker NewsI tested meshtastic in a major european city with pretty much 100% mesh coverage and its real life performance was quite underwhelming. Often I would receive messages that I could not reply to because of differences in antenna gain and crappy mesh performance. Public chat was either completely dead or flooded with test messages. Everything was super slow because the mesh can’t actually scale that well and craps out with more than a 100 nodes. Even medium fast channel would clog up fast. I would never depend on meshtastic during an emergency because it barely works even when nobody is using it. I think a public wifi mesh would be more worthwhile. Older used wifi routers are pretty much free and in unlimited supply. They use very little power. Everyone already has a compatible client device on their pocket. Sure the mesh would fail during a total blackout but at least it would be useful for something when the power is up.
Replies
Wifi routers use quite a lot of power for the area that they can cover. Ten watts or so for a hundred square meters is a lot of you want to cover a whole city.
Yeah, having gone through the blackout in Spain this would be really useful (using phones).
Then only one person needs a generator and/or Starlink to provide some connectivity.
I'm surprised that phone manufacturers haven't already implemented a mesh network. I guess you could kind of call Apple's Find My network one, but if you want to smuggle arbitary data the bandwidth is very low. Maybe Apple's new mobile Wi-Fi chip is a precursor to an actual Internet mesh network.
Meshtastic, like any other network (including mesh wifi) is not immune to the requirement for some planning.
My area has a couple of very well-placed mountaintop ROUTERs that tend to suppress most of the low level urban flooding noise, and so local messaging out to 80km or so tends to be pretty reliable. The same would be absolutely impossible with wifi.
That’s with 80 or so local nodes on LONG_FAST, population of around half a million.
Thats said, Meshtastic’s routing algorithm is extremely inefficient and has huge room for improvement.
https://en.wikipedia.org/wiki/FabFi
https://web.archive.org/web/20111119205258/http://fabfi.fabf...
I had worked with this almost 15 years ago. It was a neat project.
Meshtatisc's routing is extremely primitive and inefficient.
https://www.disk91.com/2024/technology/lora/critical-analysi...
I agree with this assessment. I've been running two nodes for about a year, maybe longer, and in that time, I've only had perhaps two contacts.
Even with a YAGI or a dedicated pole antenna, both tuned to 868 MHz, the range in my location is quite poor. The signal seems to drop off quickly, even after walking just a kilometer down the road. While I understand that height is key (and my antennas are fairly high), it appears that 868 MHz attenuates very rapidly.
So, to reiterate, I don't believe Meshtastic is a particularly effective solution. The principle behind it is sound, but the practical execution falls short. I think established methods like Hamnet and traditional amateur radio are far superior, especially now with SDRs making a simple handheld radio incredibly affordable (around €20)