- Research on fairness in voting protocols with heterogeneous weights has been published, introducing an asymptotically fair voting scheme.
- The study focuses on binary voting protocols and explores the robustness of the consensus protocol used in the next generation protocol of IOTA.
Introducing an asymptotically fair voting scheme
Recently published research delves into the topic of fairness in voting protocols with heterogeneous weights. The study presents a novel voting scheme that aims to achieve asymptotic fairness across a wide range of weight distributions. The research highlights the importance of equitable voting systems and introduces the concept of greedy sampling to improve fairness in voting outcomes.
— IOTA (@iota) May 10, 2023
Robustness of the consensus protocol in IOTA
The research project initially sought to demonstrate the robustness of the consensus protocol employed in the next generation protocol of IOTA. Splitting and merging effects, which are undesirable in a decentralized and permissionless distributed system, were the primary focus of investigation. The study references specific sources for further information on these effects.
The article concentrates on fairness in binary voting protocols, drawing inspiration from the observations made by Marquis de Condorcet on voting principles in 1785. Assuming a large population of voters, each independently casting their votes correctly with a probability greater than 1/2, the study reveals that the likelihood of the majority vote outcome being correct increases and approaches certainty as the sample size grows.
The primary motivation behind this research was to showcase the robustness of the consensus protocol used in IOTA’s next generation. Splitting and merging are two undesirable effects in decentralized and permissionless distributed systems. For more detailed insights into these effects, specific references [11, 12] are provided.
To explore fairness in voting protocols with heterogeneous weights, the research paper refers to existing literature on detecting systematic modulation of the basic Zipf law and fitting more accurate models. The study focuses on distributions that resemble the Zipf law without the need to verify certain test conditions. A visualization in Figure 1 presents the distribution of IOTA for the top 10,000 richest addresses, which aligns with a fitted Zipf law.
Based on the universality phenomenon, the plausibility of hypotheses 1) – 4) mentioned in the research, along with Figure 1, the weight distribution is assumed to follow a Zipf law when specifying a weight distribution. More precisely, for each value of n ∈ N and a parameter s > 0, the weight distribution p(n)_j is defined as 1/js * Σn_i=1(1/is) for j ≤ n, 0 for j > n (as shown in Equation 2.6). The sequence (p(n)_j) for a fixed j is decreasing in n. Additionally, when n goes to infinity, if the parameter s is strictly larger than 1, the sequence (p(n)_j) converges to a positive number, while it converges to 0 when s ≤ 1.
In summary, the published research contributes to the understanding of fairness in voting protocols by introducing an asymptotically fair voting scheme. It also sheds light on the robustness of the consensus protocol used in IOTA and explores weight distributions that resemble the Zipf law. These findings provide valuable insights for designing equitable voting systems and improving the performance of decentralized networks.