I’d say China is one of the countries which really needs dweb, but it is kinda different. Perhaps they haven’t tried DHT peer enumeration, by experience, and torrent is quite fast and not blocked here. They have developed GFW, so they will probably have the capability when they think they need to.
If we limit the nodes a peer can know and access, the network will become less open and robust. Is F2F network the only solution then ? Dweb would be way less interesting without DHT.
I remember that some people expected China not to ban the protocols because of economic concerns. They have banned cryptocurrency, private education and gaming industry. If dweb becomes a threat, they will ban it instantly as long as possible.
China is also exporting their censorship technologies I’m afraid that there will be more countries deploying these things. Optimistically, when they should have realized the threat of dweb, it would be too late, or they actually would try to ban some of it rather than everything.
Is there any solution to DHT enumeration censorship?
Yes, it’s consensus based protocols that forces everyone to agree on the state like blockchains for example.
Filecoin itself relies on DHT in the first place
That is not a hard stop, ethereum for example doesn’t use a DHT, it just do peer exchanges.
I belive the main reason filecoin use a DHT is to help find storage providers but if that ever is needed it would be easy to just publish peer records (pubkey + addresses) on the chain.
Storage providers are pretty stable usually, so you shouldn’t need to update them regularly.
IPFS can’t work with only peer exchange. Blockchains can work without DHT because peers are downloading the same thing. If you mean giving up on DHT, and the network should all store and seed the same piece of block or a DAG, the dweb will become useless due to the inefficiency. In that case, F2F protocols are better.
Btw it is really inconvenient to buy some coins in China due to recent regulations.
No, Filecoin won’t help here.
But yes, it’s possible (ish) depending on what you mean by “enumerable”. Basically, you encrypt as much data as possible:
- All DHT servers will be enumerable. The protocol requires it.
- There’s no real way to hide the fact that you’re using IPFS as long as you use a DHT. An attacker can just listen for inbound connections on the DHT to learn about all participants.
- Peers could, in theory, have “hidden” identities where finding a specific peer would be predicated on knowing the peer’s ID. This can be done by (a) deriving a secondary peer-routing key from the peer’s ID (e.g., hashing it) and (b) encrypting the peer’s peer routing record with the peer’s ID.
- Content could also be hidden(ish) to all but those that already know the CID by similarly encrypting the content routing records with the CID.
But honestly, all of these methods are best-effort. Against a state actor, they’re pretty slim. For example, if an attacker already knows the CID of some target content, hiding requests for that content in the DHT becomes impossible.
This is a fundamental issue with open systems.