"Excuse my ignorance, but isn't the overlay network setup problem one that has problems at almost every level of the stack?"
It works for me and has worked for others. The keyword here that distinguishes this idea from almost every other "peer-to-peer overlay network" project that you can read about is "small". If you limit the size of the network, you can avoid some problems. Most projects you read about aim at the ability to create a single, large network that potentially everyone can join. Open to the public. However using a different approach it is possible to create only small networks that are only open to people you know, e.g., friends, family, co-workers. There are still problems, but there are always going to be some problems. The internet you are using right now has problems. The question is does it work well enough. The small overlay network idea has worked well enough for me that I consider it one of the better ones.
It is really impossible to debate these ideas on the internet. Opinions are strong and negativity is even stronger. If you want an answer you need to try things out yourself and draw your own conslusions. No peer-to-peer solution is "perfect" and if you are always looking for the solution with zero negatives and zero limitations, you will never find it. Worse is if you never actually try these solutions, you just read about them. After you try many of them and learn what you like and do not like about the design/implementation, it is easier to chose one idea that works for your use case. Everytime someone starts promoting a peer-to-peer project you can quickly evaluate it, based on what types of designs/implementations have worked for you and which ones didn't. Well, that's my opinion, anyway.
It works for me and has worked for others. The keyword here that distinguishes this idea from almost every other "peer-to-peer overlay network" project that you can read about is "small". If you limit the size of the network, you can avoid some problems. Most projects you read about aim at the ability to create a single, large network that potentially everyone can join. Open to the public. However using a different approach it is possible to create only small networks that are only open to people you know, e.g., friends, family, co-workers. There are still problems, but there are always going to be some problems. The internet you are using right now has problems. The question is does it work well enough. The small overlay network idea has worked well enough for me that I consider it one of the better ones. It is really impossible to debate these ideas on the internet. Opinions are strong and negativity is even stronger. If you want an answer you need to try things out yourself and draw your own conslusions. No peer-to-peer solution is "perfect" and if you are always looking for the solution with zero negatives and zero limitations, you will never find it. Worse is if you never actually try these solutions, you just read about them. After you try many of them and learn what you like and do not like about the design/implementation, it is easier to chose one idea that works for your use case. Everytime someone starts promoting a peer-to-peer project you can quickly evaluate it, based on what types of designs/implementations have worked for you and which ones didn't. Well, that's my opinion, anyway.