130
2

Asynchronous Approximate Agreement with Quadratic Communication

Main:18 Pages
2 Figures
Bibliography:3 Pages
Appendix:6 Pages
Abstract

We consider an asynchronous network of nn message-sending parties, up to tt of which are byzantine. We study approximate agreement, where the parties obtain approximately equal outputs in the convex hull of their inputs. The seminal protocol of Abraham, Amit and Dolev [OPODIS '04] achieves approximate agreement in R\mathbb{R} with the optimal resilience t<n3t < \frac{n}{3} by making each party reliably broadcast its input. This takes Ω(n2)\Omega(n^2) messages per reliable broadcast, or Ω(n3)\Omega(n^3) messages in total. In this work, we present optimally resilient asynchronous approximate agreement protocols which forgo reliable broadcast and thus require communication proportional to n2n^2 instead of n3n^3. First, we achieve ω\omega-dimensional barycentric agreement with O(ωn2)\mathcal{O}(\omega n^2) small messages. Then, we achieve edge agreement in a tree of diameter DD with log2D\lceil \log_2 D \rceil iterations of a multivalued graded consensus variant for which we design an efficient protocol. This results in a O(log1ε)\mathcal{O}(\log\frac{1}{\varepsilon})-round protocol for ε\varepsilon-agreement in [0,1][0, 1] with O(n2log1ε)\mathcal{O}(n^2\log\frac{1}{\varepsilon}) messages and O(n2log1εloglog1ε)\mathcal{O}(n^2\log\frac{1}{\varepsilon}\log\log\frac{1}{\varepsilon}) bits of communication, improving over the state of the art which matches this complexity only when the inputs are all either 00 or 11. Finally, we extend our edge agreement protocol to achieve edge agreement in Z\mathbb{Z} and thus ε\varepsilon-agreement in R\mathbb{R} with quadratic communication, in O(logMε)\mathcal{O}(\log\frac{M}{\varepsilon}) rounds where MM is the maximum honest input magnitude.

View on arXiv
@article{erbes2025_2408.05495,
  title={ Asynchronous Approximate Agreement with Quadratic Communication },
  author={ Mose Mizrahi Erbes and Roger Wattenhofer },
  journal={arXiv preprint arXiv:2408.05495},
  year={ 2025 }
}
Comments on this paper