Shared-nothing Parallelization

In this pattern, different Virtual Machines (VMs) do not share any data, avoiding write conflicts such as double spending issues.

This means each VM can process transactions independently without worrying about data conflicts with other VMs.

Based on transaction preprocessing, once the simulation execution is complete, the memory areas involved in the transaction, related addresses, and other data are added to a queue, which is then assigned to different virtual machines for processing.

Shared-nothing Parallelization

The Simulation Execution module simulates the execution of transactions in the TX Pool stack, while the Onchain Operator is responsible for establishing a continuous channel to acquire data, including current block height, gas, and other information.

After this, transactions are grouped, and transactions processed by each independent VM will have no data overlap or conflict, and related transactions will be quickly completed within the same VM.