Skip to content

Latest commit

 

History

History
10 lines (6 loc) · 1 KB

TODO.md

File metadata and controls

10 lines (6 loc) · 1 KB

Roadmap:

  • Compression. Google uses an extremely fast compression algo called Snappy throughout its infra - Bigtable databases, RPC calls, etc. Calldata compression is a well-applied in the Optimism rollup.
  • Parallel execution engine. For the POC, the executor is a single node. There is a design here for a parallel executor, which uses the state dependencies of a transaction to schedule nonconflicting tx's in parallel.
  • Distributed execution network. On top of the parallel executor, we can distribute execution to a network of executors (as one executor is likely capped by it's number of CPU's), further increasing throughput.
  • Decentralized storage layer. Based on a Google Bigtable design, implemented using P2P tech. Recursive STARK proofs + data availability tech.
  • STARK-based execution proofs. We can guarantee the writes to storage are authorised by running the transaction inside a provable VM, such as Cairo.