# DAG-based Consensus with Asymmetric Trust [Extended Version]

**Authors:** Ignacio Amores-Sesar, Christian Cachin, Juan Villacis, Luca Zanolini  
**arXiv:** [2505.17891](https://arxiv.org/abs/2505.17891)  
**Date:** 2025-05-23  
**Summary page:** https://www.summarizepaper.com/en/paper/2505.17891v1/

## Key points

- In protocols with asymmetric trust, participants can make individual trust assumptions about others
- Directed Acyclic Graph (DAG)-based consensus protocols with asymmetric trust have been introduced
- The DAG-Rider protocol building blocks have been extended to the asymmetric model
- The first randomized asynchronous DAG-based consensus protocol with asymmetric quorums has been developed
- This research advances understanding and capabilities in developing consensus protocols within Directed Acyclic Graph structures

## Summary

, , , , 

In protocols with asymmetric trust, participants are free to make individual trust assumptions about others, as captured by an asymmetric quorum system. This differs from symmetric quorum systems and threshold models where all participants share the same trust assumption. While reliable broadcasts, shared-memory emulations, and binary consensus have been achieved with asymmetric quorums, this work introduces Directed Acyclic Graph (DAG)-based consensus protocols with asymmetric trust. To achieve this, the key building blocks of the DAG-Rider protocol have been extended to the asymmetric model. Surprisingly, replacing threshold quorums with their asymmetric counterparts in the existing constant-round gather protocol did not result in a sound asymmetric gather primitive. This highlights the need for new ideas in designing asymmetric DAG-based consensus protocols that rely on common-core primitives. As a response to this challenge, the first asymmetric protocol for computing a common core equivalent to that in the threshold model has been introduced. This breakthrough leads to the development of the first randomized asynchronous DAG-based consensus protocol with asymmetric quorums. This protocol is capable of deciding within an expected constant number of rounds after an input has been submitted, where the specific constant depends on the chosen quorum system. Additionally, Appendix A provides a mechanical proof demonstrating the impossibility of extending the usual gather protocol to accommodate asymmetric quorums. The model considered involves a system of processes interacting asynchronously through message exchanges, where correct processes adhere to their protocols while faulty or Byzantine processes may crash or deviate arbitrarily from their specifications. Overall, this research significantly advances our understanding and capabilities in developing consensus protocols with asymmetric trust models within Directed Acyclic Graph structures.