3 ms·
I'm trying to understand the central technical argument being made here. Please tell me if I got this right. --- Somebody has a stake in a PoS crypto currency
by perlgeek 5y ago
I'm trying to understand the central technical argument being made here. Please tell me if I got this right.
---
Somebody has a stake in a PoS crypto currency. They can now do two things: 1) sell their stake 2) sign something fraudulent (like a double spend).
Since there is no decentralized timestamp service, a node validating those two actions doesn't know how to order them, so different validating nodes come to different conclusions, and no global consensus is reached.
---
Is that what the article is trying to say?
And if yes, isn't the solution fairly easy? Within the same "chain link" of the block chain, require each action singed by the same private key to have a strictly monotonic sequence number, and if two actions appear with the same sequence number, discard both these two and all actions signed by that private key.
- ComodoHacker 5y ago> have a strictly monotonic sequence number Then you lose decentralized property.
- perlgeek 5y agoNote that I meant a strictly monotonic sequence number PER PRIVATE KEY. This only means that each holder of a private key must have some sort of synchronization mechanism (if they use several agents/clients), but it doesn't centralize the whole network.
- meribold 5y agoI think another top-level comment (by csomar) answers your question: > you can sign a transaction later and have a lower timestamp, there is nothing that prevents you from that.
- DennisP 5y agoPoS does produce a reliable monotonic sequence, it's just that it doesn't go back to genesis. Every few months you need a checkpoint hash. The idea here is that this is rare enough so it's easy to find the accepted checkpoint out of band, like we pretty much do for client software already.