Design Ideas: How should we design the logic to verify pinning (retrievability) of pinned files?

[wesfloyd]

Starting a community discussion - lemme know what you think :)

---

Draft design:

(Note: there are many, many, many potential solutions to this problem. For now, we're starting with an example high level, simplest route to achieving these goals)

• At a regular interval (eg daily) a randomized group of nodes are chosen to be act as a "retrievability" proving committee. The committee will choose a subset of CIDs to confirm retrievablity.
• The committee will attempt to connect to each IPFS server, downloaded the file, and verify it matches the CID.
  Eg ipfs swarm connect ipfs swarm connect /ip4/<server_ip_address>/tcp/4001/ipfs/<peer_id> , ipfs get <ipfs_file_hash>
• Committee updates the AVS contracts with the peerID, CID and success or failure of the download operation.