4 ms·

How does peer discovery work? Im assuming there must be some central server(s) to handle the bootstrap? I had a quick scan through the github projects, but coul

How does peer discovery work? Im assuming there must be some central server(s) to handle the bootstrap? I had a quick scan through the github projects, but couldnt see any high level documentation explaining the architecture. Would appreciate some insight!

jakswa 6y ago
"Currently searches across and advertises on the Bittorrent DHT, centralized DNS servers and Multicast DNS simultaneously."
- Sayrus 6y ago
  Implementations details can be found on discovery-channel GitHub's [0] As far as I can tell Cabal uses discovery-swarm [1] for connection management. [0] https://github.com/maxogden/discovery-channel https://github.com/maxogden/discovery-channel [1] https://github.com/mafintosh/discovery-swarm https://github.com/mafintosh/discovery-swarm
- supermatt 6y ago
  Oh yeah, I understand that bit - but how does it connect to the DHT in the first place. I couldnt see any information on where it connects when you arent yet peered (i.e. how does it bootstrap).
  - oftheoaks 6y ago
    There are a collection of hyperswarm bootstrap node IPs that can get the discovery process started for new peers.
    - supermatt 6y ago
      Many thanks - is this in the codebase or handled via some config? I couldnt see it from my (limited) searching.
nanomonkey 6y ago
It's built on the Hypercore Protocol [0](previously known as DAT), so I would assume DHT (Distributed Hash Tables). [0]https://hypercore-protocol.org/ https://hypercore-protocol.org/