The replication factor is determined by means of checking the entire blockchain for the similarity of tags? Who is doing this?

III. CONSENSUS: Each farmer now stores a provably unique replica of the ledger, and we define the replication factor as the number of unique ledger replicas stored across the network.

The replication factor is determined by means of checking the entire blockchain for the similarity of tags?
Who is doing this?

2 Likes

Good question!

There are a few things that allow us to do that:

  • slot probability: farmers are solving on each time slot (once a second), but only one actual solution that allows to produce a block is expected to be found every 6 time slots (once every 6 seconds)
  • solution range: range of solutions in the plot that are considered valid from consensus point of view, similar in nature to work difficulty in PoW chains
  • piece size: size of one piece of the archival history that farmer plots to disk

By observing the quality of the solutions on the network and taking into account the size of one piece, we can estimate how much space should have been plotted by the network in order to find just one solution with given slot probability. This is similar to the estimation of the total hash power in PoW chains based on the observed work difficulty.

Here are the exact lines of code that do the estimation in the current codebase: