sub0 Asia 2024 - Kamil Salakhiev (Quadrivium)
summary
Kagome C++ client is already syncing with Kusama, running a validator on Westend and undergoing a security audit to add a second Polkadot host implementation. Current relay-chain blocks come mostly from the Aura fallback, creating predictable leaders; the plan replaces probabilistic BABE-plus-Aura with Sasafras ring VRF tickets so exactly one anonymous leader produces each slot, reducing forks and enabling cache and parachain performance gains. Implementation targets include asynchronous backing with fragment trees, elastic sharding across cores, and a GD grid topology for statement distribution.
AI-written summary · about
chapters (16)
- 0:002024 Polkadot changes
- 1:05Polkadot Host and runtime
- 2:08Good block production engine
- 3:11BABE in block production
- 4:39Secondary block production issues
- 5:22Sassafras block protocol
- 7:51Sassafras ticket assignment
- 9:37Sassafras benefits and limits
- 10:43Sassafras caching optimizations
- 12:10Parachains protocol security
- 14:55Synchronous backing
- 16:18Asynchronous backing throughput
- 17:45Fragment trees in backing
- 18:52Elastic scaling for parachains
- 19:12Grid topology for statements
- 21:19Conclusion and outlook
related talks
27mPolkadot Behind the Code, Episode 2 and Capturing Behind the Code Panel
22mBoundless
23mWhat If... Kubernetes Core Pieces Could Be Extended with Pluggable WebAssembly Modules?
14mIntroducing Kroma: Road to EVM-Equivalent ZK Rollup
37mDistributed Flexibility: Nomad and Vault in a Post-Kubernetes World
21mCOSMWASM Powerups!
21mFireside Chat
56mJAM & DOT DAOism