I always treated "blockchain finality" as closer to marketing language than a real guarantee, somewhere between hope and probability. Reading how Dusk's Succinct Attestation actually closes a block changed that.
Each round runs through three phases: a randomly selected provisioner proposes a candidate block, one committee votes on its validity, a second committee ratifies that outcome, both using aggregated BLS signatures reaching a supermajority. That's not a probabilistic confirmation stacking up over time, it's an explicit cryptographic attestation the block satisfies. Blocks move through defined states, attested, confirmed, final, and DuskDS settles in around ten seconds, DuskEVM even faster at roughly two. Oak Security audited the whole consensus and economic protocol and called it well-designed, combining pieces of existing approaches with some genuinely custom ones.
That's a real answer to reorg risk. A trade that's final on Dusk isn't final until nobody objects for a while, it's final because a defined committee already attested it is, which is the actual requirement for regulated securities settlement.
Here's what it doesn't resolve, and it connects to something I raised weeks ago about DuskEVM's sequencer. That attestation process is fully decentralized among provisioners on DuskDS. But DuskEVM still orders transactions through a single sequencer before anything reaches DuskDS for that same guarantee. The finality I just described protects what happens after ordering. It was never built to answer who sees the transaction first.
#dusk $DUSK @Dusk #DuskEVM
Each round runs through three phases: a randomly selected provisioner proposes a candidate block, one committee votes on its validity, a second committee ratifies that outcome, both using aggregated BLS signatures reaching a supermajority. That's not a probabilistic confirmation stacking up over time, it's an explicit cryptographic attestation the block satisfies. Blocks move through defined states, attested, confirmed, final, and DuskDS settles in around ten seconds, DuskEVM even faster at roughly two. Oak Security audited the whole consensus and economic protocol and called it well-designed, combining pieces of existing approaches with some genuinely custom ones.
That's a real answer to reorg risk. A trade that's final on Dusk isn't final until nobody objects for a while, it's final because a defined committee already attested it is, which is the actual requirement for regulated securities settlement.
Here's what it doesn't resolve, and it connects to something I raised weeks ago about DuskEVM's sequencer. That attestation process is fully decentralized among provisioners on DuskDS. But DuskEVM still orders transactions through a single sequencer before anything reaches DuskDS for that same guarantee. The finality I just described protects what happens after ordering. It was never built to answer who sees the transaction first.
#dusk $DUSK @Dusk #DuskEVM
