DNSLink lets a regular domain name resolve to IPFS content using nothing but a DNS TXT record. Here is the exact syntax and how it compares to IPNS.

IPNS solves the one thing a CID can't: staying the same address while the content behind it changes. Here is how a name gets derived, signed, and updated.

IPFS content verification lets a client check retrieved data against its CID directly, without trusting whichever peer or gateway sent it. Here is how.

IPFS libp2p is the peer-to-peer networking stack that DHT lookups and Bitswap transfers actually run on. Here is what it handles: identity, security, and NAT.

The raw IPFS Merkle DAG has no concept of a file or folder. UnixFS is the format that adds one. Here is exactly how, node type by node type.

Unpinning an IPFS file does not delete it immediately. Here is exactly what pin types exist, what garbage collection actually removes, and when.

Content addressing vs location addressing is the core design split between IPFS and HTTP. It fixes tampering and link rot, but creates a new problem: discovery.

IPFS splits content retrieval into two jobs: a DHT finds who has your file, Bitswap fetches it. Here is exactly how provider records and want-lists work.

IPFS defaults to 256 KiB fixed-size chunks, but also ships Rabin and Buzhash content-defined chunkers. Here is why the choice affects deduplication.

An IPFS CID is not a random string. It encodes a version, a codec, and a hash algorithm plus digest. Here is how to decode a real CID piece by piece.

IPFS does not use a B-tree for large directories. It uses a HAMT, a Hash Array Mapped Trie. Here is exactly how it shards a folder once it outgrows one block.

How does IPFS work under the hood? Every file becomes a Merkle DAG of hash-linked blocks, found via a DHT and moved with Bitswap. Here is the full path.