A move to create a .p2p top-level domain outside of the regular DNS root is under way.
Following the outcry over the US government’s seizure of 82 .com domain names this weekend, a group of coders have decided to create a namespace not overseen by ICANN (which had nothing to do with it).
It’s not entirely clear to me how many projects have launched.
(Interestingly, dot-p2p.org appears to have been registered several days prior to the weekend’s domain name seizures)
The .p2p project plans to create an application that would intercept all DNS requests for .p2p domains and route them via a peer-to-peer network rather than the user’s regular DNS servers.
This presumably means that the entire .p2p zone file could wind up being stored on endpoints, which sounds like a scalability challenge to me.
More problematic is the the issue of “decentralization”, which is of course critical when you’re talking about trustworthy DNS. It can be summed up in this sentence:
“Hello, I’m bankofamerica.com.”
If anybody can claim to own any domain name, you need to be able to figure out who’s telling the truth.
The .p2p initiative seems to be dealing with this by, um, centralizing control over .p2p domain assignments to a free “registrar” at nic.p2p.
To prevent warehousing, registrants would need to prove they already own the string in another TLD in order to register the equivalent .p2p domain.
The project is obviously in its very early stages, as demonstrated by this wiki page, which tries to figure out the problem of decentralization using some kind of trust/voting system.
Here’s an example of the lack of thought that seems to have gone into it so far:
A small conflict, not malicious
1. Alice assigns fbi.p2p -> 220.127.116.11.
2. Bob propagates the assignment to his node, because he trusts Alice.
3. Dave assigns fbi.p2p -> 18.104.22.168. Conflict created.
4. Carol sees the conflict and:
– Decides to just follow the decision of her trustees and assigns fbi.p2p -> 22.214.171.124, or
– Does not create any assignment. There will be a warning and she will try to work out the problem with others.
5. Everyone will try to agree on a solution.
The page also currently includes this beauty:
1. Chuck owns a botnet and uses 10^6+ zombies to game the system.
2. Shitload of fake request need to be disproved
4. Problem? :U
The project seems like a heck of a lot of wheel-reinventing in order to solve a problem that doesn’t exist.