3 ms·
IPFS is a collection of many different protocols put together. So regarding naming, the modularity of IPFS gives way for having multiple different naming-schema
by diggan 8y ago
IPFS is a collection of many different protocols put together. So regarding naming, the modularity of IPFS gives way for having multiple different naming-schemas pointing to the same content. And users can verify this, as it's all content-addressed.
This image is usually very helpful to look at to understand the architecture: https://ipfs.io/ipfs/QmXzrZY68r958KZtaWMo7DHRB29dePhxSv7DYFm28etqSb/ipfs-layers.jpg https://ipfs.io/ipfs/QmXzrZY68r958KZtaWMo7DHRB29dePhxSv7DYFm...
So one way to archive human-readable names is having a DNS record pointing to a hash. So the user would still go to `example.com` but underneath, it would be served `/ipfs/QmaSD`. It could also use IPNS (currently build-in naming system for PeerIDs in IPFS) to point to the same content.
Once there are more (or better) naming systems, it'll be easy to add them to IPFS.
But currently, DNS is probably the most used together with IPNS, as most people are used to DNS and it generally works (but not offline, so far from perfect)
Edit: re your own example, with DNS + IPFS it would go something like this: you ask "What the content of the records of example.com?" DNS resolves answers: "This Gateway + This Hash" and you request the hash from that gateway.
I actually made a video explaining the whole dnslink thing as well if you're interested in some visualization on how it works: https://www.youtube.com/watch?v=YxKZFeDvcBs https://www.youtube.com/watch?v=YxKZFeDvcBs