was rereading section 7 of the dusk whitepaper last night ands got stuck on something i hadnt clocked before.
most people talk about proof of stake like theres one role, you stake, you validate, done. dusk actually splits that into two separate jobs.
generators propose the next block, provisioners validate and finalize it. different responsibilities, different extraction methods.. different stake requirements even.
what's interesting is generators get selectd through a privacy.
preserving procedure so the leader for a given round isnt known in advance. provisioners get pulled into committees through a separate process entirely. two roles means two places a attacker would need to control stake in, not one.
hadnt thought about consensus design as a role separation problem before this, mostly just thought 0f it as stake more get picked more.
does splitting proposer and validator roles.
actually raise the cost of attacking a chain, or is it mostly an. organizational choice??
#dusk @Dusk $DUSK
most people talk about proof of stake like theres one role, you stake, you validate, done. dusk actually splits that into two separate jobs.
generators propose the next block, provisioners validate and finalize it. different responsibilities, different extraction methods.. different stake requirements even.
what's interesting is generators get selectd through a privacy.
preserving procedure so the leader for a given round isnt known in advance. provisioners get pulled into committees through a separate process entirely. two roles means two places a attacker would need to control stake in, not one.
hadnt thought about consensus design as a role separation problem before this, mostly just thought 0f it as stake more get picked more.
does splitting proposer and validator roles.
actually raise the cost of attacking a chain, or is it mostly an. organizational choice??
#dusk @Dusk $DUSK