Bachelor/Master Thesis

Asymmetric Zoo

Distributed systems are prone to failures. It is therefore necessary to make assumptions about the possible failures that might occur to be able to solve problems in these environments. In most systems, all participants must have the same assumptions, also referred to as symmetric systems. Asymmetric trust is a novel idea that allows participants in distributed systems to use their own failure assumptions, independent from the choices made by other processes. This more accurately models the real world, where different people have different trust notions. These ideas have also been adapted by companies into real systems [8, 9, 10] worth more than 130 billion dollars [11].

There have been many proposals for a model of asymmetric trust [1, 2, 3, 4, 5, 6, 7]. All of them have in common the idea of allowing different participants to chose different trust assumptions, however, they also differ significantly in some more concrete details. Some of them allow the system to not have global quorum intersection [2, 6], others require global quorum intersection but not necessarily in correct processes [4, 5], while others require quorum intersection in correct processes [7]. There are also differences in the way in which processes are characterized, some models characterize processes using a finite and small number of labels [4, 6], while others have proposed a more fine-grained classification of processes [5]. All these differences make it hard to compare and integrate the results obtained in the different models.

In this thesis we will explore and systematize all models, we will find equivalencies between them, explore if there are models that contain others, and, if possible, try to derive a more general model that encompasses all of them.

References

[1] Heterogeneous Paxos

[2] The Stellar Consensus Protocol

[3] Secure Protocols with Asymmetric Trust

[4] Asymmetric Distributed Trust

[5] Weaker Assumptions for Asymmetric Trust

[6] Stellar Consensus by Instantiation

[7] Quorum Subsumption for Heterogenous Quorum Systems

[8] Stellar

[9] XRP Ledger

[10] Anoma

[11] CoinMarketCap

Contact Juan Villacis for more information.

Nature of the project: Theory 100%.