Researchers Develop Signature-Free BFT Consensus for Low-Latency Systems

Simple-IT, a new signature-free Byzantine fault-tolerant (BFT) consensus protocol, offers a theoretical latency of 4 message delays, only one more than the optimum, and key latency of 3 on its optimistic path. This simplicity enables practical implementation and competitive throughput and latency, as demonstrated in tests. Qianyu Yu of The Hong Kong University of Science and Technology and colleagues from University of Bern, Stellar Development Foundation and Aptos Labs, have created Simple-IT, a new system designed to protect distributed computer networks from attacks enabled by future quantum computers.

The protocol offers an alternative to costly security upgrades by removing the need for complex digital signatures, instead relying on secure communication channels. Most existing systems use digital signatures to verify message authenticity, but these are susceptible to attacks from future quantum computers. Specifically, algorithms like Ed25519 and BLS12-381, commonly employed in BFT protocols, are vulnerable to Shor’s algorithm, a quantum algorithm capable of breaking the underlying discrete logarithm problem upon which their security relies. Simple-IT offers a solution by removing the need for these signatures, instead utilising secure communication channels, and achieves a theoretical latency of four message delays, only slightly above the ideal minimum. The team demonstrated competitive throughput and latency in tests, but the vital question remains whether this simplified approach can truly match the performance of established, yet vulnerable, protocols, and offer a flexible path to post-quantum security.

Reduced latency and increased throughput in a novel Byzantine fault-tolerant consensus protocol

Simple-IT, a new signature-free Byzantine fault-tolerant (BFT) consensus protocol, achieves a theoretical latency of 4 message delays, exceeding previous complex protocols without practical implementations. It represents a key improvement, being only one message delay greater than the theoretical optimum for such systems, and attaining 3 delays on its optimistic path. The theoretical minimum latency for BFT consensus is generally considered to be 3 message delays, achievable only under ideal conditions and with specific assumptions about network synchrony. Previously, low-latency signature-free protocols remained largely theoretical due to implementation complexities and the difficulty of achieving consensus without relying on cryptographic assurances. These complexities often stemmed from the need for intricate message ordering and validation mechanisms to compensate for the absence of signatures.

Researchers at University of Bern and The Hong Kong University demonstrated that Simple-IT’s streamlined design enables competitive throughput, peaking at 170,000 transactions per second, with latency comparable to existing quantum-vulnerable protocols. Recent advances in quantum computing pose a looming threat to most current Byzantine fault-tolerant (BFT) consensus protocols, which rely on quantum-vulnerable public-key signature schemes such as Ed25519 and BLS12-381. An alternative to switching to more expensive post-quantum secure signature schemes is the use of signature-free protocols, relying only on cheap, post-quantum secure authenticated channels. This investigation explores whether signature-free BFT consensus protocols can match the performance of current state-of-the-art, quantum-vulnerable BFT consensus protocols. The significance of this lies in the potential to mitigate the quantum threat without incurring the substantial computational overhead associated with post-quantum cryptography, which often involves significantly larger key sizes and more complex operations.

Prior to work on the Sailfish++ protocol showed that state-of-the-art throughput is attainable without signatures, but the question of latency remained open. Several recent signature-free protocols have low latency in theory, yet lack practical implementations. Simple-IT, a new leader-based, signature-free BFT consensus protocol, achieves a theoretical latency of four message delays, or three on its optimistic path. Its simplicity allows for practical optimizations like speculative pipelining, where nodes can pre-process subsequent blocks before receiving final confirmation, thereby reducing overall latency. Experimental results in a geo-distributed testbed demonstrate performance competitive with state-of-the-art quantum-vulnerable protocols. These results confirm that signature-free, leader-based BFT protocols can effectively rival the speed of established systems. However, current figures are based on controlled network conditions and do not yet reflect performance in highly adversarial or unpredictable real-world deployments. The protocol’s success depends on utilising authenticated channels, secure communication links verifying identity, to sidestep the need for costly upgrades to post-quantum cryptography, offering a viable pathway towards strengthening distributed systems against the anticipated threat of quantum computers compromising current cryptographic methods. The leader-based approach involves a designated node proposing blocks, which are then validated by other participants, simplifying the consensus process but introducing the potential for leader failure or malicious behaviour, which the protocol is designed to tolerate.

Performance evaluation is limited by undisclosed network characteristics

Distributed systems face an ever-present battle against potential disruption, and Simple-IT offers a compelling response to the looming threat of quantum computers breaking existing security measures. The authors openly admit a vital detail limiting broad assessment of their findings: the precise configuration of their geo-distributed testbed remains undisclosed. Gauging reliable performance outside this controlled environment is difficult without knowing the specific network parameters, bandwidth, latency variations, or node distribution. Factors such as network topology, geographical distance between nodes, and the presence of network congestion can significantly impact performance, and without this information, it is difficult to extrapolate the results to other settings.

Despite this limitation, the demonstration of comparable performance with existing systems remains significant. This approach establishes that signature-free BFT consensus can rival the performance of current systems vulnerable to quantum computing attacks; BFT protocols ensure agreement in a network even when some participants fail or act maliciously. The protocol achieves this by employing a multi-phase commit process, where nodes exchange messages to reach a consensus on the order of transactions. The team’s work highlights the potential for streamlined designs to deliver substantial gains in distributed system performance and security. Further research is needed to assess the protocol’s resilience in more realistic and challenging network conditions, including scenarios with high levels of network latency, packet loss, and malicious actors. Investigating the scalability of Simple-IT to larger networks and its integration with existing distributed systems are also important areas for future work.

Simple-IT demonstrated that a signature-free Byzantine fault-tolerant consensus protocol can achieve performance competitive with current state-of-the-art systems. This is important because existing protocols rely on public-key signature schemes vulnerable to advances in quantum computing. The protocol achieves a theoretical latency of 4 message delays and 3 on its optimistic path, offering a viable alternative without switching to more expensive security measures. Researchers suggest further work is needed to assess its resilience and scalability in more complex network environments.