Is it possible to create a mesh 5G network run by volunteers supplying their connectivity, thus remove the need for mobile network operators?
The biggest hurdle is that reliably running high performance transmitters is not easy for amateurs, and the payoff for any one transmitter is not that much. I'm going to use the example of a residential ISP but this applies to cell networks as well. The "meshier" the network is, the more people revenue needs to be split between, exacerbating the problem.
Another issue is that reliability is extremely important for internet access. Given the fact that amateurs are not going to be able to maintain high uptime, for a decentralized mesh network to succeed at actually providing internet service, you need to have a lot of redundancy in any given area, further reducing income from any one node.
The solution to this is to have a team of technicians that can go around and fix and optimize nodes as soon as there is any problem. This is basically what an ISP or cell carrier does. An added difference is that in a mesh network, the idea is generally that the property owner owns the node, while with a conventional ISP, the property owner leases to the ISP who owns the node. Property owners generally prefer the latter, since this is the model they are used to operating under as landlords.
An end user wants to pay a steady monthly fee for internet that never goes down. A property owner wants to get a steady monthly check for leasing a site. A technician wants to get a steady salary for fixing nodes.
A decentralized repair network is likely to reduce reliability for end users, predictability for property owners, and job security for technicians. All three of these parties may find it more optimal to have an ISP business which can finance and coordinate things, even if the business is taking profit which could have gone to the other participants.
You'd just need one crooked technician who can recruit some number of node owners. (If the tech did it via remote-controlled power interruption and only did it on scattered nodes in areas with lots of redundancy, they wouldn't even need to recruit node owners and split the cash. But they'd also be leaving evidence that could easily get them in trouble once someone started getting wise to it.)
Take a look at WiFi-dense apartment buildings. So much crowding, no centralized assignment or management of the bands. It is a wild west of people transmitting on whatever channels and whatever power levels they want (within the legal limits). It ends up with few people actually having a good experience when there's no centralized management. 5GHz/6GHz makes WiFi more usable because it naturally limits your ability to hear your neighbors. Going to 700MHz/900MHz/1.2GHz (the normal frequencies used in a lot of 5G deployments) is only going the opposite direction of where WiFi has been going to solve this problem. Expect more noisy neighbor problems as you lower the frequencies.
Then we're not only going to saturate the bands with people doing whatever they want (within legal limits), we're going to depend on mesh routing through all that noise? There goes your reliability and efficiency of sending data.
And who's to say they want to join your mesh and not Bob's super awesome mesh? Or start their own mesh? Oh, you get to decide how to operate the mesh but I can't? I guess you'll end up getting some kind of license so you can standardize how this particular mesh should operate and prevent others from running competing services on the same frequencies as your one mesh.
You'll put out standards on what kinds of devices are certified to work on it and ensure certain settings so tx/rx errors are reduced to ensure good usage. You'll start encouraging people to not put up more nodes in a certain area because it's just getting too crowded here, but hey we need to incentivize someone to set up a node on the other side of town.
Snap now it seems like we're running a regular carrier.
I participate and use city-sized WiFi mesh networks in the amateur radio world. They're not anywhere near a replacement for what normal people think of as internet connectivity. I can't imagine swapping WiFi for 5G cellular stacks would end up making a radical difference. The issues are largely with having to make multiple wireless hops, mesh routing inefficiencies/problems, and having everyone actually play nice all the time.
Why would it be part of every modem/gateway? Since there's no monetary incentive to participate, in all likelihood all nodes would be run by volunteers who are shelling out extra for a compatible modem/router.
Actually come to think of it, you can run a volunteer network providing internet connectivity with off the shelf equipment right now. It's called setting your wifi network to "open". Why don't people do that? How would your mesh network fix those issues?
That's not really a "mesh" then. It is just a bunch of infrastructure AP's everywhere.
If it's a part of every modem/gateway router, why would you bother routing it through a bunch of mesh hops just to eventually get out instead of just routing it through the far more reliable wired networking available at every modem/gateway router?
Those regular WiFi networks only have tons of available bandwidth because they're not trying to repeat a bunch of wireless traffic. Even the current mesh WiFi networks only really work when you're using frequencies that aren't trying to compete with neighbors. Start getting actual density and it'll all fall apart.
Also your idea of "standard QoS can still apply" isn't exactly true. That QoS is only going to work if people play along with it. In the end its a shared medium. Get some clients to not play along with your configurations, you'll start getting collisions regardless of what you configure your QoS settings.
The question is can devices connect to other devices that route to router with internet. Is it possible to have router for house without internet connection that routes to the neighbors that do?
The bandwidth you care about in this scenario, where you don't have backhaul, is in the air. And very limited.
My neighbor still has the Helium antenna and radio on his balcony but it's offline due to costs/profits disproportion. It's the LoRaWAN, pre-5G hardware though, and I don't know anyone running the 5G version, if it's even a real thing. I liked the idea from technical perspective but the project itself was off-putting for me due to being built around a crypto token and having overall web3 smell.
One, what frequency are you going to use? If you use 2.4GHz or 5GHz, your Wifi-using neighbors will hate you. There is the 6GHz spectrum but has problems with long ranges. The 3.5GHz CBRS is probably the best bet but that requires spectrum allocation and organization to run it. The mobile operators have all the good low frequency, long range spectrum.
Two, the range with home routers is going to be pretty short, maybe 1mi. That means lots of node to cover a city. Also, 5G routers are not that cheap. It also means that there will be no reception away from the city. Most routers are meant to be used inside, and good coverage, requires mounting them outside on a pole.
Three, I'm not sure there is 5G device-to-device. There was LTE-Direct but it never got implemented. There D2D in 5G spec but I can't find any implementations.
However, creating the Wikipedia/Internet Archive of wireless ISPs would be pretty awesome.
Not used everywhere, but seen as something that would be rolled out for critical communications, natural disasters, etc.
So, if you want to run legally, you're going to need spectrum licenses and transmitter licenses and all that. That will make you a mobile network operator, regardless of how you arrange labor and sites.
change my view
Mesh would be each home (or some percentage of the homes) act as nodes. These have all the homes hit a few towers around the city. Traffic isn't routed directly between (or through) the homes in this example, it is all centralized. They hit a single big tower that then does all the routing.
(It doesn't help that people in this thread are confusing "mesh" with "collection of access points each of which is individually connected to a wired ISP", which is not a mesh.)