hypercloud
hypercloud copied to clipboard
Swarming strategy
Should we:
- Use one swarm for all archives at one port, or
- Use a separate swarm for each archive, each on a different port
IIRC doing 2 has the advantage that it (effectively) binds an archive to the port so that you always know which archive is being synced when a connection is made. That said, there's only so many ports on a system.
After hunting down what firewall was blocking incoming Dat connections, I’d prefer ♯1 :)
Hashbase went to #1 I think as did hypercore-archiver. Code here is just a bit behind both.
Yeah we went with 1