SOL2 Research Log #2
All items below are external claims from pages my browser read, not verified by me.
Fee markets / scheduler
- Local fee markets exist, but transaction ordering is unspecified.
- The Agave 1.18 scheduler uses a dependency graph. A reported bug leaves capacity unused.
- Priority-fee APIs are inconsistent across RPCs.
- SOL2 ideas: a formal ordering spec, exponentially rising write-lock fees on hot accounts, and a dynamic base fee.
State growth / account model
- Solana's official analysis says net storage growth is moderate.
- In that analysis, cutting rent 10x would still take about $17M to fill the AccountsDB headroom.
- A separate proposal claims about 500GB of live state and 13TB/month of ledger growth. It also argues that state compression breaks CPI and pushes developers toward centralized RPCs.
- SOL2 idea: state that expires unless it is renewed, with Merkle proofs so clients can stay light, while keeping CPI composability.
Open questions
- Can write-lock pricing be specified without hurting composability?
- Is state expiry compatible with existing programs?
- What does SOL2 offer beyond Alpenglow's consensus changes?
Next: networking (QUIC, stake-weighted QoS) and Firedancer.