Briefing
The foundational problem in Byzantine Fault Tolerance (BFT) consensus is the unrealistic reliance on a symmetric quorum system, where all participants must share the same global trust assumptions. This research introduces the first asymmetric DAG-based consensus protocol, which fundamentally re-engineers the core mechanisms of DAG-Rider to accommodate locally defined quorums , allowing each node to specify its own trust set based on its unique information. The breakthrough lies in formulating a new asymmetric protocol for computing the common core primitive, a necessary component for transaction ordering, which is proven to be functionally equivalent to its symmetric counterpart. This new theoretical model enables randomized asynchronous consensus that decides within an expected constant number of rounds, a critical implication for unlocking high-performance, low-latency BFT in truly heterogeneous and decentralized blockchain architectures.
Context
Established BFT consensus protocols, including seminal works like PBFT and HotStuff, are architected upon the assumption of symmetric quorums. This theoretical constraint dictates that every participant in the network must agree on the identical threshold of trusted nodes necessary for safety and liveness. In practice, this means the consensus mechanism is rigid, failing to account for the real-world heterogeneity of decentralized networks where nodes operate across different jurisdictions, possess unique social connections, or have access to divergent local trust information. This limitation prevents BFT from achieving its full potential in permissionless or highly diverse environments.
Analysis
The core mechanism is the extension of the Directed Acyclic Graph (DAG) consensus model to an asymmetric trust framework. The paper replaces the standard symmetric quorum with asymmetric quorums , which are locally defined by each participant to reflect their individual trust choices. The critical innovation is the development of a modified gather protocol that specifically handles these divergent views to compute a common core → the block of transactions that all honest nodes agree to order. This asymmetric common core protocol is not a mechanical substitution of the standard quorums; it includes additional steps that ensure the DAG structure can still enforce a consistent, global transaction order despite the heterogeneous trust inputs, thus preserving the necessary safety and liveness properties of BFT.
Parameters
- Expected Constant Rounds → The time-to-finality for any input is an expected constant number of rounds. This metric is the asymptotic lower bound for latency in asynchronous BFT systems.
- Total Participants to Smallest Quorum Size Ratio → This ratio is the specific factor that determines the magnitude of the expected constant latency. A smaller ratio implies faster finality.
Outlook
This research opens a new, fundamental avenue for designing BFT protocols that are resilient and flexible enough for real-world deployment in highly decentralized settings. The theoretical success in achieving constant-time finality under an asymmetric trust model suggests that future blockchain architectures, particularly those utilizing DAGs for concurrency, can move beyond the artificial constraints of global trust assumptions. In the next three to five years, this work is expected to enable new generations of asynchronous, high-throughput Layer 1 and Layer 2 sequencing protocols that can scale to a massive number of heterogeneous nodes without sacrificing security or performance.
Verdict
This theoretical breakthrough fundamentally re-engineers the trust model for BFT, proving that high-performance, constant-time consensus is achievable without the restrictive assumption of a globally shared, symmetric quorum.
