Industry Research Report: Latest Developments and Applications of Blockchain Consensus Protocols

Source: Matthieu Rambaud, Télécom Paris

Adapted by: Anonymous, Coin World

Report Summary

This report explores the current development status of blockchain consensus protocols, with a focus on the latest advancements in Asynchronous Byzantine Fault Tolerance State Machine Replication (BFT SMR) protocols. The fastest asynchronous protocol currently is 2-chain VABA, but due to its vulnerabilities, its expected delay of 9.5δ has not been achieved. Therefore, sMVBA has become the fastest asynchronous MVBA protocol, with an expected delay of 10δ. The report also introduces two new protocol designs, namely 2PAC (2-Phase Asynchronous Consensus) and Ultra-Fast Pipelined Blocks, demonstrating significant improvements in throughput and latency.

Blockchain technology, as a decentralized distributed ledger technology, ensures the integrity and consistency of data through consensus mechanisms. The consensus mechanism is the core of the blockchain system, and its performance directly affects the scalability and security of the blockchain. Asynchronous Byzantine Fault Tolerance (BFT) consensus mechanisms have unique advantages in dealing with network delays and partial node failures, making them a key area of research.

Models and Definitions

In the asynchronous BFT model, the system consists of n = 3f + 1 processes, where f processes may be maliciously compromised by adversaries. These processes communicate with each other through asynchronous channels, with message transmission delays controlled by the adversary. Each process has a pair of public and private keys for signing and verifying messages, ensuring the authenticity and integrity of the messages.

Blockchain Consensus

Blockchain consensus protocols aim to achieve agreement among all honest nodes regarding the state of the blockchain. Specifically, each node continuously receives new transactions and packages them into blocks, ensuring that these blocks reach consensus among all honest nodes through the consensus protocol. Blockchain consensus protocols must meet the following basic requirements:

Liveness: In infinite executions, there exists an infinitely long decided blockchain.

Consistency: If there are two decided blockchains, one must be a prefix of the other.

P-Quality: In the decided blockchain, the proportion of transactions input by honest nodes must be at least p.

Current Challenges in Asynchronous Consensus Protocols

The fastest current asynchronous consensus protocol is 2-chain VABA, with an expected delay of 9.5δ. However, we have identified various attack vectors that compromise its consistency and liveness. For example, attacks due to lack of verification checks, attacks that hinder liveness using elevation strategies, and consistency attacks resulting from relaxed leader authentication definitions. Although 2-chain VABA introduces some new mechanisms, such as running multiple parallel instances, it still fails to completely resolve these issues.

New Protocol Design: 2PAC (2-Phase Asynchronous Consensus)

Based on an analysis of existing protocols, we propose the 2PAC protocol. This protocol significantly improves performance by simplifying and optimizing the consensus process. Specifically, it includes two variants:

2PAClean:

Achieves +90% throughput and an expected delay of 9.5δ, with message complexity of O(n²).

Improves the efficiency of the protocol by eliminating unnecessary interactions and computational overhead.

2PACBIG:

Currently the fastest blockchain consensus protocol with message complexity of O(n³).

The fault-free single MVBA run time is 4δ, greatly reducing latency.

Ultra-Fast Pipelined Blocks

We propose a new pipelined block design that significantly reduces the latency of pipelined blocks. By introducing a fast path mechanism, under a fair scheduler, the decision time for pipelined blocks is even smaller than that of non-pipelined blocks. This mechanism guarantees the latency of the fast path in all executions and is unaffected by the behavior of faulty processes.

Quantitative Results

Through theoretical analysis and practical testing, 2PAClean has an expected delay of 9.5δ in the worst case, while in the good case (no faults and semi-fair scheduler), it is 6δ. In contrast, the expected delay of sMVBA is 10δ, and in the good case, it is 6δ. Therefore, 2PAClean reduces the worst-case delay by 0.5δ while maintaining the same good-case delay. Additionally, the throughput of 2PAClean is improved by 80% to 100% compared to chain-based sMVBA, primarily due to the new design avoiding unnecessary block discards and computational overhead.

As a protocol with message complexity of O(n³), 2PACBIG has a single MVBA run time of 4δ, faster than all existing protocols. Furthermore, the ultra-fast pipelined block design allows s2PAClean and s2PACBIG to achieve decision times of 4δ and 3δ for pipelined blocks, respectively, further enhancing the performance of the protocol.

Computational Evaluation

To verify the performance of the new protocols, we conducted extensive computational evaluations. The results show that 2PAClean and 2PACBIG exhibit excellent performance under various network conditions, particularly in high-latency and high-failure-rate environments. Specifically, 2PAClean achieves a good balance between message transmission latency and computational complexity, while 2PACBIG realizes lower latency through parallelization and optimization of the voting process.

As blockchain technology continues to evolve, asynchronous BFT consensus protocols will play an increasingly important role in ensuring security and improving performance. The designs of 2PAC and ultra-fast pipelined blocks illustrate the future direction of blockchain consensus protocol development, which aims to achieve higher throughput and lower latency through simplified protocol structures and optimized consensus processes.

Future research can further explore the following directions:

Protocol Optimization: Further simplify and optimize protocol structures to reduce unnecessary message passing and computational overhead.

Security Analysis: Conduct in-depth analyses of the new protocols' security in various attack scenarios to ensure their reliability in practical applications.

Practical Applications: Apply the new protocols to actual blockchain systems to validate their performance in real network environments.

This report provides a detailed analysis of the strengths and weaknesses of current asynchronous blockchain consensus protocols and proposes two new protocol designs, namely 2PAC and ultra-fast pipelined blocks. The new designs demonstrate significant advantages in improving throughput and reducing latency, providing important references for the future development of blockchain technology. These new protocols not only theoretically prove their superiority but also showcase outstanding performance in practical testing, offering new ideas for achieving efficient and secure blockchain consensus protocols.

Through continuous research and optimization, we have reason to believe that blockchain technology will play an increasingly important role in the future digital economy, and the next generation of consensus protocols will provide a solid foundation for the development of this technology.