Veriblock Architecture for Blockchain Transaction Throughput
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Blockchain networks face scalability issues with transaction throughput due to the complexity of consensus execution and messaging overhead, especially as the number of nodes increases, limiting the ability to process transactions efficiently and securely.
Innovation Solution
The Veriblock architecture introduces an execute-order-validate model that uses a hybrid approach combining Verifiable Computing (VC) and Trusted Execution Environment (TEE) to execute and verify smart contracts, allowing for parallel transaction processing and reducing the number of endorsers required, thereby enhancing transaction latency and security.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the number of nodes (replicas) in the blockchain network is increased to improve security and decentralization, then the network becomes more secure and distributed, but the transaction throughput decreases due to increased consensus complexity and messaging overhead
Solution Approach 1:
The patent divides the blockchain network into two distinct sets of nodes: committing replicas (a small set of n nodes) that perform consensus and commit transactions, and endorsing replicas (a larger set of N nodes) that validate transactions through gossip protocols. This segmentation allows the network to maintain high security through a small trusted consensus group while achieving high throughput through parallel validation by many endorsing nodes, thus resolving the contradiction between network security and transaction throughput.
2Reliability
If traditional BFT consensus is used to ensure security and consistency, then the network achieves strong security guarantees, but the messaging complexity and computation overhead increase significantly, slowing down transaction processing
Solution Approach 1:
The patent extracts the heavy consensus execution burden from all endorsing nodes and concentrates it solely on the small set of committing replicas. Endorsing nodes perform lightweight validation through gossip protocols without executing full consensus, while committing replicas handle the computationally intensive BFT consensus and transaction commits. This extraction significantly reduces transaction latency for the majority of nodes while maintaining security through the dedicated committing replica group.
3Reliability
If endorsements are replicated multiple times across many nodes to ensure validity, then the consistency and security of transactions are improved, but the transaction throughput is reduced due to the overhead of multiple endorsement replications
Solution Approach 1:
The patent segments the endorsement validation process by having endorsing nodes perform lightweight validation and propagate endorsements through gossip protocols, while the actual consensus validation is performed only by the small set of committing replicas. This allows many endorsements to be generated in parallel by numerous endorsing nodes without requiring all of them to participate in the heavy consensus process, thus maintaining transaction consistency while improving throughput.
Data Source
Figure 1A
Figure 1B
Figure 2A
AI summary
Transaction executions/commits in a blockchain network need to be fast, robust and secure and thus calls for minimal latency in transaction commits. In an execute-order-commit blockchain network, latency is high due to smart contracts been executed at every endorsing node of the blockchain network. A method and system for processing transactions in the blockchain network is disclosed. The system discloses a veriblock architecture, which enables processing a transaction request by executing an associated smart contract along with a proof of correctness of execution of smart contract using only one endorser. Further, enables verifying the smart contract by multiple endorsers. The smart contract associated with the proof, referred herein as a vericontract, is executed to generate an output and the proof using one of a) Verifiable Computing (VC) approach, b) a TEE approach and c) a hybrid approach (combination of VC and TEE).