| |
| ▲ | perching_aix 2 days ago | parent | next [-] | | Yes, they just don't track the individual torrents anymore. They only play a role during the initial peer discovery stage (bootstrapping). Peers find torrent swarms on their own, the bootstrap servers are excluded from all that. | |
| ▲ | LtdJorge 2 days ago | parent | prev [-] | | If you are connected to the DHT network, you don’t. Unless you mean for DNS and such. | | |
| ▲ | perching_aix 2 days ago | parent [-] | | No, they mean specifically that in order to connect to such a network, you need to hit some specific central nodes first. | | |
| ▲ | LtdJorge 2 days ago | parent [-] | | The bootstrap nodes. But those don't _need_ to be centralized, even if they have historically been to some degree. There coul be millions. | | |
| ▲ | perching_aix 2 days ago | parent | next [-] | | You still need to have a list of a subset of them, and you just have to trust they connect you to the "same network", to the extent that even makes sense, no? | |
| ▲ | immibis 2 days ago | parent | prev [-] | | If there are millions of bootstrap nodes, how do you find them? | | |
| ▲ | toast0 2 days ago | parent | next [-] | | If they listen on a well known port, and there are millions, send out a few thousand probes to 'random' IPv4 addresses and you'll most likely find one. If you get and keep a list of bootstrap nodes when you find one, then you can random select from the bootstrap addresses rather than all routable IPv4 addresses. | | |
| ▲ | immibis 2 days ago | parent [-] | | What's your IPv6 plan? | | |
| ▲ | toast0 2 days ago | parent [-] | | Probing IPv6 would be pretty difficult. You could make some assumptions that might hold, like assume only the lowest /64 out of a /48 is used and ::0 or ::1 for the end of the address. Likely you'll still need too many probes to be feasible. You'd need to probe v4 space through NAT64 and exchange v6 addresses after that, or include a cache of viable v6 addresses with clients. That gets you close to centralized service again, because how do you get the viable addresses to distribute with the client? Probably by running a supernode and dumping the list of supernodes into the client source every so often; but starting off with just that node listed. |
|
| |
| ▲ | LtdJorge 2 days ago | parent | prev [-] | | You can embed them in the torrent, same as the tracker. But the tracker has to handle all the state, while a bootstrap node just hooks you up with a few (or one) node in the DHT network. After that, it's fully decentralized. There could other ways, embedding them on SRV DNS records, etc. It's the same issue as with getting a DNS server. You could in theory get addresses of bootstrap nodes from your ISP through DHCP (lol, sure). |
|
|
|
|
|
