|
|
|
|
|
|
by jarym
23 days ago
|
|
|
Iroh can use multiple discovery methods, one of them is DNS. When a node comes online it publishes its addresses[1]. Another node can resolve from the node id those addresses and attempt a direct or hole-punched connection. If that works, relays aren't needed. So agree not serverless, but also does not NEED a coordination server (unless you class DNS as a coordination server which is debatable). [1] https://docs.iroh.computer/concepts/address-lookup#endpoint-... |
|
|