In September 2022 I gave a talk at code.talks in Hamburg, “Decentralized Storage with IPFS: How Does It Work Under the Hood?”. It is an introduction for developers. IPFS is easy to use from the command line, so the talk looks at what happens below the commands.
It covers four topics:
- Content addressing and content identifiers (CIDs).
- How files are split into chunks and linked in a Merkle DAG with IPLD.
- The Kademlia DHT, which spreads provider records over the network with XOR distance and bucketed routing tables.
- The full flow of publishing and retrieving content: bootstrapping, building the routing table, and storing provider records on the 20 closest peers.
The same data fed into our SIGCOMM ‘22 paper, Design and Evaluation of IPFS. I gave the prize talk for that paper a year later, see Applied Networking Research Prize, IETF 118 Prague.
The recording is on YouTube.