Comment by api
4 years ago
How large can these networks be in practice?
I know a major issue with most mesh protocols is that when the network is both big and dynamic configuration traffic will start to swamp data traffic.
Another issue in most mesh protocols is susceptibility to Sybil, flooding, cache pollution, and other denial of service attacks by well-resourced adversaries. It's really hard to create a distributed permissionless system that can't easily be attacked in those sorts of ways. Proof of work and similar schemes are really the only ways we've found so far and those are expensive and cumbersome.
Any mesh network used by dissidents is going to be targeted for denial of service attacks.
Even ignoring the concept of adversaries, wireless mesh is difficult to scale. Throughput and range are inversely proportional. Environmental conditions can attenuate signal. Propagation of routes is tricky as the network grows. It makes sense for something like starlink where there is LOS and ownership. But for modern day bandwidth requirements wireless mesh doesn’t make sense. Useful for smaller deployments where the data payloads are small and the max hops is limited.