I've heard it said that if you stand up a new DHT it will be subject to Sybil attack until it gets very very large, and so the only reasonable DHT is Mainline DHT since it is already large enough to withstand this.

This is true for a Kademlia DHT. But with newer DHT research it is no longer true (for example, Whanau). And there has been about 20 years of research since Kademlia, and the Internet has changed too (NAT is even more of a problem now), plus I have a few other reservations like mutable data must be very small, disappears rather quickly, isn't fixed into the protocol yet (many nodes don't participate) and it also feels like an abuse of a DHT that was meant for bittorrent (just like inscriptions/ordinals/NFTs feels like an abuse of bitcoin).

I think it is worth researching DHT technology further, but of course mainline DHT might still be the right answer, I'm just not convinced given the concerns I've mentioned.

Nonetheless, pkarr is pretty fucking cool. The rest of pubky looks to me like an attempt to do something interesting with pkarr.

nostr:npub1jvxvaufrwtwj79s90n79fuxmm9pntk94rd8zwderdvqv4dcclnvs9s7yqz

Reply to this note

Please Login to reply.

Discussion

You might be interested on hash function Blake3, which enables parallels Processing and native Torrenting and realtime streaming:

https://en.m.wikipedia.org/wiki/BLAKE_(hash_function)#BLAKE3

"In addition to providing parallelism, the Merkle tree format also allows for verified streaming (on-the-fly verifying) and incremental updates.[27]"

"BLAKE3 is a single algorithm with many desirable features (parallelism, XOF, KDF, PRF and MAC), in contrast to BLAKE and BLAKE2, which are algorithm families with multiple variants. BLAKE3 has a binary tree structure, so it supports a practically unlimited degree of parallelism (both SIMD and multithreading) given long enough input. The official Rust and C implementations[27] are dual-licensed as public domain (CC0) and the Apache License.[28]"

Yes I am a fan of BLAKE3 and I would pull for it if we started over with a nostr2. But a DHT is *far* more than a hashing function. And existing DHTs have already defined their hashing function (mainline uses SHA-1 since it started long ago).

Worth more research, a Nostr web-of-trust might be a way to mitigate Sybil attacks.