Principle of hashing DHT chunks?

Question

bittorrent

Hello everyone, I can’t figure out the essence of how clients work with DHT, namely, how chunks are hashed depending on each other for one file and how the user looks for them inside the table? I also had 2 additional questions:
1) Does the algorithm connect chunks of one torrent file "geographically" in DHT? Well, let's say I divided the file into several pieces, will their hashes be similar or completely independent? If they are similar, then how does it turn out, then they have different bytes. What algorithm does this?
2) If they are independent, how will they be searched, that is, the client will start searching for the nearest nodes by their hashes separately? That is, specifically, it turned out 3 hashes: 870b, aaa2, efed A node client that needs these files will look for them from completely different providers, independently of each other? Such an operation should slow down the download compared to the tracker system, because the complexity of finding log (n) (for example, in Kademlia DHT), it turns out that for these 3 "non-neighbor" hashes, you need to go through the nodes 3 times, while the tracker almost immediately will give Roy with the data. In practice, it's slower than the tracker, or am I wrong?

Answer 1

torrent dht is a protocol for distributing not pieces of a torrent, but actually the torrent file itself, with identification by its hash and, of course, for obtaining a list of peers for downloading it.
the first article in Russian about dht

Answer 2

DHTused only to search for seeds. It is not used when transferring files.