In case you haven't yet heard the great news, libp2p (opens new window) can now punch holes.
This is a companion discussion topic for the original entry at https://blog.ipfs.io/2022-01-20-libp2p-hole-punching/
This is a companion discussion topic for the original entry at https://blog.ipfs.io/2022-01-20-libp2p-hole-punching/
1 Like
Thank you for this helpful and clear article.
Is there any place where I can find a fully working example of that, w/ go-libp2p ?
1 Like
Do these examples help at all? go-libp2p should more-or-less figure out the hole-punching for you with little intervention. Lots of upgrades / changes / techniques happen over time too, which all make it into the spec and libraries.
Does the data plane pass through the relay mode being used?
Iām assuming yes because thatās the purpose of bypassing the firewall.
Iām asking about traffic engineering here.
Thatās to say, I may be taking a suboptimal route via the relay node.
Also, is it possible to āinterceptā connections via relays in a firewall.
Thoughts?
Iāve asked similar questions but there doesnāt seem to be much interest in running IPFS on networks that you donāt control. I personally donāt like the, āWTF are you doing?ā email from the IT department.
1 Like
Relays are used to coordinate the hole punching protocol, theyāre not doing any other kind of data-relaying by default.
OK thanks, so the hole is punched simply to gain access to the metadata locating the peers. Once both have that, they can make outbound connections with the right IPās and port numbers?