Anyone here an expert on DHTs, or maybe an enthusiast? Back in 2018 when I was designing the Intercoin Protocol (which we now call the Intercloud) I met with P

Anyone here an expert on DHTs, or maybe an enthusiast? Back in 2018 when I was designing the Intercoin Protocol (which we now call the Intercloud) I met with Petar Maymounkov, who invented Kademlia in the early 2000s. Dude used to teach at NYU (my old school) so I emailed him and took him to Olive Tree Cafe — wound up asking him a bunch of questions and learning a lot. Anyway, I have a question about all this. Consensus is an expensive operation and th reason Blockchains are slow is because they run fu network consensus cor every block. In Intercloud, we run network-wide consensus only occasionally, and mostly to update a provably random seed (by combining input from many different nodes). This serves as a seed for a random oracle for the network. Now, that oracle is used during “shuffling”. I think we may have originally got the idea from the SAFE project. What happens is that, every so often, swarms of nodes are shuffled and nodes are reassigned to other swarms. So although they may pick their original swarm, they have no control over what future swarms they will be in, because it is effectively impossible for then to predict, let alone control, the random oracle and hashes of activity from neighboring swarms. Being off by a single bit throws off any of their well-laid plans. Shuffling is necessary to break up swarms that have been corrupted, overrun by dishonest nodes sybil attacks, or just unable to reach consensus due to unavailability. So most of these DHT issues go away. However, encryption is still a very good idea. Another GREAT idea is what SAFE pioneered - removing the IP addresses from responses after the first hop. Because otherwise it is trivial to DDOS a network or a subset of it. PS: last year I interviewed Ian Clarke (creator of Freenet) about this stuff https://m.youtube.com/watch?v=JWrRqUkJpMQ https://m.youtube.com/watch?v=JWrRqUkJpMQ