[Consensus] Write-up rationale for VDFs in EC

sternhenri commented 5 years ago

What it is This could take a number of formats (GH comment below, doc, etc), but it simply seeks to rationalize in a single artefact why EC needs VDFs. This includes arguments about:

PoS replication of PoW guarantees (specifically for SPC)
Potential attacks that are mitigated by VDF

Why it matters This is a question we keep discussing, having this para could help generally onboard or level people up/track our thinking.

sternhenri commented 5 years ago

@ZenGround0 feel free to close if you no longer feel this is useful.

ZenGround0 commented 5 years ago

Hey I know this is closed which is fine by me as I don't think this is launch blocking. But for more info on this topic see this issue in the archived aq repo.

sternhenri commented 5 years ago

Thanks @ZenGround0 I closed it for the reason you point out and the fact I think we have adequately aligned as a research team on why we want VDFs. For quick summary (I will provide longer write up in a forthcoming paper):

VDFs allow us to:

Get rid of clock synchrony assumption that most consensus algorithms make, by effectively enforcing clock synchrony through waiting.
Prevent folks from simulating the chain in the past (useful for a series of attacks), aside from the entropy provided by multiple participants on the chain
Likewise prevent ticket grinding in running leader election, ensuring that elections remain power based rather than compute based (as one might by grinding through tickets in parallel)
Prevent miners from rushing the protocol, thereby hurting miners who have worst connections and impacting protocol fairness.

filecoin-project / consensus

[Consensus] Write-up rationale for VDFs in EC #25