Omakase framework for parallel execution in multi-proposer blockchains

Omakase uses conflict-graph integration to improve transaction execution throughput.

Omakase is a parallel execution framework. It combines pre-execution, inter-block conflict-graph integration, and graph-aware Optimistic Concurrency Control (OCC).

Replicas use idle cores to pre-execute only the blocks they propose. This spreads prediction work across concurrent proposers.

A greedy scheduler integrates conflict graphs across blocks committed in a round. It keeps an integration only when a lightweight balance heuristic predicts a parallelism gain.

A graph-aware OCC executor runs the integrated graph. It admits transactions in dependency order to eliminate aborts when the predicted graph is exact.

Omakase reached a 2.47x2.47\text{x} speedup over sequential execution. This was tested on 15,000 Ethereum mainnet blocks and synthetic workloads.

Omakase matches mean throughput against the strongest alternative. It is more stable under high contention because integration bounds optimistic work in flight.

Key facts

  • Omakase reaches a 2.47x2.47\text{x} speedup over sequential execution.
  • Testing used 15,000 Ethereum mainnet blocks containing approximately 2.4M transactions.
  • Integration hides 94% of its measured cost in the validator pipeline.

All sections

Edition