Briefing
The prevailing limitation in DAG-based Byzantine Fault Tolerance (BFT) protocols is their reliance on a symmetric trust model, where all nodes share the same fault tolerance assumptions. This research introduces the first asymmetric protocol for computing a “common core” primitive, the foundational building block for commit rules in DAG consensus. This new primitive enables the construction of a randomized asynchronous DAG-based consensus protocol that operates effectively with asymmetric quorums, where each process maintains unique trust assumptions. The single most important implication is the creation of highly robust, heterogeneously-trusted decentralized networks that can achieve consensus with an expected constant number of rounds, significantly expanding the real-world applicability of high-throughput DAG architectures.
Context
Before this work, the established theory for DAG-based consensus, exemplified by protocols like DAG-Rider and Bullshark, required a “gather” protocol or similar common core primitive to ensure a shared set of vertices (state) among all correct processes. Critically, these primitives were only proven correct under the assumption of symmetric quorums, meaning the system’s fault tolerance was uniformly defined. This imposed a theoretical limitation on building systems where individual nodes might have different, verifiable trust relationships with their peers, forcing a single, global trust parameter that limited flexibility and resilience.
Analysis
The core mechanism is the asymmetric common core primitive. Previous symmetric protocols relied on the Q3-condition for quorum systems, which assumes a uniform failure model. The new logic extends this by defining and proving a common core that functions with asymmetric quorums, where each node maintains its own local view of trusted peers (its quorum).
The protocol uses randomization to ensure that despite the heterogeneous trust landscape, all honest nodes can still agree on a common set of vertices in the DAG within a bounded, expected number of communication rounds. This fundamentally decouples the system’s liveness and safety guarantees from the restrictive requirement of a universally agreed-upon, symmetric trust threshold.
Parameters
- Finality Latency → Expected constant number of rounds – The protocol guarantees that finality is reached within a predictable, bounded number of communication steps, regardless of the network’s size.
- Trust Model → Asymmetric Quorums – Each node defines its own set of trusted peers, allowing for heterogeneous security assumptions across the network.
- Core Primitive → Asymmetric Common Core – A new foundational component that replaces the symmetric “gather” protocol in DAG-based BFT.
Outlook
This foundational work opens new research avenues in designing decentralized autonomous organizations (DAOs) and cross-chain communication protocols that operate under a flexible, heterogeneous trust landscape. In the next 3-5 years, this theory could unlock real-world applications in federated identity systems and supply chain management, where participants naturally have asymmetric trust relationships. It provides the cryptographic and distributed systems primitive necessary to move beyond the rigid, uniform trust models that currently constrain enterprise and consortium blockchain adoption.
Verdict
This research fundamentally redefines the trust model for DAG-based consensus, establishing a new, robust primitive for building decentralized systems with heterogeneous security assumptions.
