#dusk $DUSK @Dusk

When I first loked at Dusk, I was mostly focused on Phoenix, staking and the privacy side. But the more I dig into it, the more Kadcast catches my attention.

Dusk uses Kadcast as its P2P broadcast layer, and it’s basically designed to make sure a blok doesn’t depend on just one perfect network path.

The interesting part is the redundancy. Kadcast can send the same block chunk to multiple delegates instead of trusting one node. The protocol uses a configurable β value, and the implementation supports β = 3, so if one path fails or a node is unavailable, other paths can still carry the data forward. It also uses Forward Error Corection (FEC) because UDP can lose packets. Research on Kadcast shows this combination can keep propagation reliable even with packet loss.

And this maters for Dusk because blocks are limited to 1 MB, roughly 250 Phoenix transactions. Dusk says Kadcast can reduce bandwidth usage by around 25–50% compared with common gossip aproaches.

I like this design because fault tolerance here isn’t only about consensus. It starts with making sure the message actually gets around.

Do you think better block propagation is underrated in blockchain security?