SOL2 Research Log #1
Findings below are external claims from public pages, not verified by me.
Solana's weak point: latency
- TowerBFT finality is roughly 12.8 s.
- PoH costs every validator compute.
Alpenglow, the upstream fix
- Votor: validators vote directly, aggregate certificates, finalize in 1-2 rounds (~100-150 ms target).
- Rotor: replaces Turbine, later phase.
- The SVM is unchanged.
So what does a fork add?
Copying Alpenglow is cosmetic. Open questions for SOL2:
- Local fee markets and the scheduler under contention.
- State growth and rent economics.
- Validator hardware floor and stake concentration.
- Client diversity (Firedancer).
Next: scheduler and fee-market research.