Browsing IPFS using Cloudflare’s gateway requires two things: a browser connected to the Internet, and the address of something on IPFS that you want to view.
As mentioned on the introduction page, every file added to the IPFS network is given a unique address based on its contents which is called a Content Identifier, or CID. So if you have an image stored on IPFS, its CID would be based on the hash of the bits that compose that image.
The job of an IPFS Gateway like Cloudflare’s is to take requests for CIDs and return the content corresponding to the given CID. Each time you access a piece of content through a gateway, you provide a URL with two parts:
cloudflare-ipfs.com, while the makers of IPFS run a gateway at
ipfs.io, and if you run your own node you may even have a gateway at
/ipns/<domain>. If the request starts with
/ipfs/, that tells the gateway that you want the content with the CID that immediately follows. Because the content is addressed by CID, the gateway’s response is immutable and will never change. If the request starts with
/ipns/, that tells the gateway that you want it to lookup the CID associated with a given domain in DNS and then serve whatever content corresponds to the CID it happens to find. Because DNS can change over time, so will the gateway’s response.
That’s generally going to look like:
the gateway that you’re providing the address of a piece of content stored on
QmXoypizjW3WknFiJnKLwHCnL72vedxjQkDDP1mXWo6uco is the address
Put those two components all together and the URL you provide will look like this:
The first link is a mirror of Wikipedia, and I can say that with confidence because it is an /ipfs/ link and therefore immutable. The second link is IPFS’s marketing site, and they can update it at any time by modifying the DNS records associated with the ipfs.io domain.