Direct swarming with the Consumer's node

rsksmart / rif-storage-pinner

MIT License

2 stars 3 forks source link

Direct swarming with the Consumer's node #64

Closed AuHau closed 3 years ago

AuHau commented 4 years ago

In order to speed up the pinning of the data it is desired to so call directly swarm with the Consumer's node. This directly connects the Provider's node to the Consumer's node and bypass the slow content-discovery of DHT.

There is also initial implementation of GraphSync (BitSwap successor) in the latest go-ipfs which can speed up the data retrival.

For this to work though, the Provider's node needs to know the Consumer's node PeerId, for which will be needed the #62

vojtechsimetka commented 4 years ago

Potential issues:

for the connection we may need to get a multiaddress from a PeerId (consider if it makes sense for the consumer to give it)

AuHau commented 4 years ago

You can find multiaddresses from DHT using ipfs dht findpeer:

USAGE
  ipfs dht findpeer <peerID>... - Find the multiaddresses associated with a Peer ID.

SYNOPSIS
  ipfs dht findpeer [--verbose | -v] [--] <peerID>...

ARGUMENTS

  <peerID>... - The ID of the peer to search for.

OPTIONS

  -v, --verbose  bool - Print extra information.

DESCRIPTION

  Outputs a list of newline-delimited multiaddresses.

With those, you can try all the public multiaddresses to swarm with.

AuHau commented 3 years ago

This was closed in #185