Introduction
Content-addressed storage represents a fundamental shift in how we think about data location. Instead of asking "where is this data?" we ask "what is this data?" The answer is a cryptographic hash that uniquely identifies the content itself.
How Content Addressing Works
In a location-addressed system, you retrieve data by knowing its address (URL, file path). In a content-addressed system, you retrieve data by knowing its hash.
Traditional: GET /files/report.pdf
Content-addressed: GET /QmXoypizjW3WknFiJnKLwHCnL72vedxjQkDDP1mXWo6uco
The hash is derived from the content itself, so:
- Immutability — if the content changes, the hash changes
- Deduplication — identical content is stored once regardless of how many times it's referenced
- Verifiability — anyone can verify they received the correct data by recomputing the hash
Git: The Most Successful Content-Addressed System
Git was one of the earliest and most successful content-addressed systems. Every commit, tree, and blob in Git is identified by its SHA-1 hash. This is what makes Git's distributed model possible:
- Branches are just pointers to commit hashes
- Merging is possible because content is immutable
- You can verify the integrity of your entire repository at any time
IPFS and the Distributed Web
The InterPlanetary File System (IPFS) extends content addressing to the web. Instead of downloading from a specific server, IPFS retrieves content from the peer-to-peer network using content identifiers (CIDs).
IPFS uses a Merkle DAG (Directed Acyclic Graph) structure where:
- Files are split into chunks
- Each chunk is hashed
- Chunks are assembled into a tree structure
- The root hash becomes the file's identifier
Building With Content Addressing
Content-addressed storage enables powerful architectural patterns:
- Supply chain provenance – track materials from source to finished product
- Decentralized identity – self-sovereign identity anchored to content hashes
- Reproducible builds – pin build inputs by content hash for deterministic outputs
- Data portability – move data between providers without losing references