Feat request: Create a policy/guideline for trusted third party volunteer public relays

[PrinceGarth]

I would look towards Tor's policies for exit node and relay operators for inspiration

• Transparent disclosure of threat model with third party public relays
• Clear privacy policies
  • Commitment from the relays not to record logs or IP addresses, etc
• Technical enforcement
  • Rate limits for connections, number of clients to relay: limits how many clients/IP addresses a malicious relay would be able to scrape
• Operator tiers instead of binary trust. New volunteer relays are to some extent untrusted (maybe rate limited?) and have to earn trust to become higher tiered.
• Method for users to report misbehaving nodes
• Who curates the list? The con of this idea is that it adds some amount of centralization, but that's already there if it currently uses bootstrap relay nodes that are hard coded, and the relays are a fallback anyway, in case peer-to-peer holepunching fails
• E2E encryption already protects the content from untrusted relays, but relays can monitor IP addresses and timing (so can ISPs if it's peer-to-peer)

This would be only complementary to the upcoming ability to use one's own trusted and reachable peers as relays. Or even made irrelevant, as some people may host public relay peers that can be added via a website (hopefully they have at least a privacy policy)

[pymq]

Thanks for initiating this. The IPFS project (the creators of libp2p) should already have similar guidelines, and we can use them as inspiration too. But for now, I think the project is still small enough that we can manage this manually.