#dusk $DUSK @Dusk

I was looking into Dusk’s csensus and one thing I didn’t expect was how much the network layer matters when you’re trying to make blocks final really fast.

A block can be perfectly valid, but if sme provisioners receive it late, the whole consensus process can lose time.

That’s where Kadcast caught my attention. Dusk uses it for message propagation, while Succinct Attestation handles proposal, validation and ratification. So block propagation isn’t jst “send it to everyone”, it has to move through the network quickly enough for committees to work on the same information.

I also noticed Dusk requires provisioners to keep nodes synchronized, with Kadcast using UDP port 9000. Even the official troubleshooting guide tells operators to check peer connectivity and block height when chain progress stalls.

And the minimum direct stake is 1,000 DUSK, meaning running consensus infrastructure isn’t just about holding tokens, the node has to stay online and synced.

For me, this is the interesting part of Dusk: fast finality is not only a consensus formula. Network propagation is part of the speed limit too.

what happens when Dusk scales to much more geographically distributed provisioners?

Could propagation become the real bottleneck?

Propagation Limit?