Open shazow opened 8 years ago
andrewtj
commented
8 years ago “Domain registration metadata” is a little vague but glue, delegation and whois data is not considered public data by all the parties involved (state, business or otherwise). (I know that might sound odd but it is what it is.)
Can you elaborate on why you think this data would be useful in “bridging traditional DNS-land with IPNS-land”?
shazow
commented
8 years ago “Domain registration metadata” is a little vague but glue, delegation and whois data is not considered public data by all the parties involved (state, business or otherwise). (I know that might sound odd but it is what it is.)
Any references to this? I'd love to read up more on it.
My only experience is looking into getting a domain dump to make a domain searcher.
Can you elaborate on why you think this data would be useful in “bridging traditional DNS-land with IPNS-land”?
I'm imagining there will be a class of clients that only get to interact with the internet over IPFS (like browser clients), so it would be valuable to lookup domains -> nameservers -> DNS entries in some way. Honestly this scenario is a bit contrived, I haven't thought about it a ton.
Also I feel this falls inline with IPFS's attitude towards getting large datasets that are (or should be) public and make them public in this way. Or maybe it's more of a thing for the IA team or something?
jbenet
commented
8 years ago @shazow solid idea-- would love to help with this, as long as the legality concerns are resolved.
@davidar maybe this could be a type of https://github.com/ipfs/archives effort?
andrewtj
commented
8 years ago On Sun, Nov 22, 2015 at 11:06 AM, Andrey Petrov notifications@github.com wrote:
“Domain registration metadata” is a little vague but glue, delegation and whois data is not considered public data by all the parties involved (state, business or otherwise). (I know that might sound odd but it is what it is.)
Any references to this? I'd love to read up more on it.
My only experience is looking into getting a domain dump to make a domain searcher.
Sorry I don't have any references for you. It's a bit of a pain to Google the topic. DENIC and NSEC enabling zone-enumeration that conflicts with German privacy law is the only concrete example that comes to mind at the minute. If you look into ICANN's revisiting of the whois protocol that'll probably lead to similar issues.
Can you elaborate on why you think this data would be useful in “bridging traditional DNS-land with IPNS-land”?
I'm imagining there will be a class of clients that only get to interact with the internet over IPFS (like browser clients), so it would be valuable to lookup domains -> nameservers -> DNS entries in some way. Honestly this scenario is a bit contrived, I thought about it a ton.
Seamless DNS integration interests me. I'm not sure what that'd look like exactly, except that I'd expect it to be part of the existing global namespace.
Problem
Domain registration metadata changes fairly rarely (few times per year for most domains), but getting an up-to-date dump of the entire set of registered domains is a massive ordeal even though this is technically public data.
Most gTLD owners require you to fill out forms, fax them, and generally get special approval before getting access to an FTP server with big collated dumps. It's not clear why there's a huge barrier to access for this data: Competitive advantage? Legacy technology? Bandwidth costs?
Solution
IPFS could host all domain registration metadata, ideally owned/provided by the gTLD owners. To boot strap this, we'd need to come up with a reasonable format/layout for storing this data on the network and tools for keeping it up to date against legacy formats. Then, we can provide initial uploads of popular gTLDs (.com, etc) before approaching the owners to take over the maintenance of this data.
Applications