Latency for content look up (retrieval market) -

Hi, I am doing in detailed study of file coin… and currently trying to understand flow of retrieval market (https://filecoin-project.github.io/specs/#systems__filecoin_markets__retrieval_market) works.

Content lookup happens with a function FindProviders(), in which a client will pass CID of piece, and the function returns “peers” who provides the content.

After going through different file coin material/videos, my understanding is: FindProviders() gets the information directly from “chain”. Each Node’s Ledger in File coin chain contains information about piece providers. So content lookup can happen in single round trip with minimal latency as each Node has “piece” index. And content look up does not involve DHT.

Is my assumption correct?

Hi @venugopv,

All current CIDs being stored and their status can be found in the latest state trie on chain. So yes, see this Slack msg: https://filecoinproject.slack.com/archives/CEHHJNJS3/p1585159620012300

To make it easier we will be proposing a Filecoin dev grant RFP for a CID tracker website / service run alongside a full node that can query / list CIDs in the latest state tree e.g. 1 long page that shows a table = each miner x sectors x state.

It can be a lookup oracle for storage status. Perhaps nodes will also implement FindProviders if lookup is efficient.