Closed dob closed 7 years ago
Yeah that's a good idea.
The code spike has 2 different commands - kad.go
is simpler to get started.
It's actually quite similar to the libp2p example, which I recommend playing around with.
p.s. livepeerd.go
is a slightly more advanced example with simple video streaming working.
This issue seems to be done. I think we should consider a better interface so it can be easily integrated and stay generic.
This is going to be a lot easier to build and debug/test if we can make use of some basic underlying P2P network, where you can assume you have at least one peer to exchange these messages with. Should we use Eric's libp2p spike as the underlying network as a starting point? We could basically import it into this project, and then swap it out as we make progress on the network layer specifically.
This way as you fire up two or three nodes locally it would be easy to add the message types, and see debug output as nodes send messages, receive messages, respond to messages, etc. They'll all just connect to each other. Or do you have another proposal for building the actual network formation from the ground up in response to the
peers
messages?