Scalable Block Execution via Parallel Validation

Abstract

A dominant part in blockchain networks is reaching an agreement on block transactions and their impact on the network state. We follow a common scenario where a node is selected to propose a block and its implied state updates. The proposal is then validated by other nodes that examine the block impact on the state. Typically, all validators execute the complete block and provide an indication based on comparing the results of their execution to the updated state in the proposal. With the increase in the number of participants in blockchain networks, we suggest a time-efficient block validation through splitting it into multiple disjoint tasks. This can be challenging due to possible dependencies between the block transactions. We describe the additional information the leader has to provide to enable that. Moreover, we describe a unique proof for the block partition computed by the leader such that when validated in part by the different committees guarantees the correctness of the execution by the leader. We compare the approach to traditional solutions based on real data of the Ethereum blockchain.