- goal is async backing is 6s. unclear if proposal will start with 12s. Parameter can be used to then tweak this. Ideally we'll propose 6s.
- launch on kusama end of september.
- We're effectively doubling the amount of cores usage.
- Load testing on versi by the end of the month (august).
- Share PRs?
- https://github.com/paritytech/cumulus/pull/2300
- https://github.com/paritytech/polkadot/pull/5022
- 500ms to 2s for a collator to build a block (4x). Then from 12s to 6s, do (8x). In 12s we would occupy half of many cores, therefore entire network throughput stays the same, but for parachains is the same.
- we'll do 6s for parachains.
- Number of cores is our bottle neck.
- We'll never have less than 6s latency. But this could be done on the collator side. Could be very very low latency. We hope to be able to group these ~100ms submissions into 6s parablocks.
- Rococo launch planned for early september. Fast transition from Rococo to Kusama.
- working on a guide for updating collators.
- If a parachain team doesn't do anything, their chain continues to work regardless.
- They need to upgrade both collator and runtime.
- Use relay chain blocktime.
- Share system chain team? Good example for the ecosystem.
- Make cumulus template by default use relay-chain blocktime.