The content is the address, no central issuer of the tracker file needed just the CID content ID. Also since the CID is a hash of the data it can be used to validate the data :wink:
That is exactly how bittorrent works as well tho. Bittorrent hasn't needed any trackers for public torrents since 2005 and can work solely off magnet links using the DHT network. Just like how IPFS does it.
the difference is, when you have a lot of (small) files.
With ipfs you have one global peer pool, with torrents you need a pool per file.
That's basically the same as torrents with magnet links.
As other commenters pointed out, the biggest real difference is that IPFS objects can cross-reference each other by hash, which allows you to do partial updates.
An “update” means a new CID though, and this basically works like a persistent data structure [0]?
That's quite literally how bittorrent with DHT works. To download a public torrent you only need its content hash, the metadata and peers are then fetched via DHT assuming at least one reachable DHT node is aware of the torrent.
Winny (since 2002) and successor Japanese P2P network were like that. They use MD5/SHA file hash as address. https://en.wikipedia.org/wiki/Winny
The biggest thing is that in IPFS each chunk has a CID, and that CID is globally unique and sharable from other CIDs.
You can think of each CID as a mangent link, and each CID can point to more CIDs.
Where as in a torrent the peer tracking is done at the whole torrent level, instead of the chunks inside the torrent.
Torrent V2[0], which has poor adoption thus far, should also allow for a similar ability. To my knowledge this is not being taken advantage of yet.
> each chunk has a CID
An extremely interesting problem, fingerprinting generic binary - reducing kilobytes and megabytes into a handful of bytes, and avoiding collision.
Much simpler on large data.
Seems like over the last 5 years they haven’t really done anything but add crypto buzzwords to the project site.