← All researchMF-PRISM-DLT-2026-04 / interactive guide

THE IDEA, MADE VISIBLE

Six validators can share one failure.

Group validators by an infrastructure domain, then remove entire domains instead of independent machines.

Research reportFault-Domain-Aware Federated Quorum Certification

Calculated illustration · change the inputs to inspect the mechanism.

01 / 04Guided chapter

Count machines

Six validator processes may look like six independent units of resilience. Their deployment relationships determine whether that count is meaningful.

Distinct hosting domains

Failed domains

Validators: 6
Survivors: 4
4-of-6 available?: Yes

Losing 1 of 3 equally sized hosting domains leaves 4 validators. The chosen availability predicate survives this scenario.

Correlated failures act on domains, not independent validator counts.

TRY THIS

Choose one domain and fail it. All six validators disappear together, even though the machine count never changed.

What this experiment represents. Constructed hosting topology and availability check only. These are not actual Metriq infrastructure records, a full fault-domain policy search or a Byzantine-safety certificate.

Source manuscript & release ↗Read the full explanation ↓

JavaScript is off. The initial figure above and the complete written explanation below remain available. Interactive controls require JavaScript.

1. Count machines

Six validator processes may look like six independent units of resilience. Their deployment relationships determine whether that count is meaningful.

2. Reveal correlated domains

Here validators are evenly distributed across the selected number of hosting domains. A domain event removes every validator assigned to it.

3. Apply an explicit fault scenario

Fail the first one or two domains and count surviving validators. The 4-of-6 availability predicate is then evaluated on that exact scenario.

4. Separate scenario checks from certification

The manuscript searches adversary policies and domain-cover witnesses. Passing one hosting scenario is not a certificate for every organization, region or correlated infrastructure event.

Distributed Systems · Research Paper

Fault-Domain-Aware Federated Quorum Certification

Adversary policies, domain-cover margins, and exact portfolio certification

Research reportMF-PRISM-DLT-2026-04 / v0.2

Certify quorum safety under correlated failures rather than validator counts alone. Organization, region and infrastructure incidents are modeled explicitly through domain-cover policies and exact witness searches.

Read the paper Release & files

Cover of Fault-Domain-Aware Federated Quorum Certification — Current public editionv0.2 · 2026-09-12

Identifier: MF-PRISM-DLT-2026-04
Series: Distributed Systems
Edition: Version 0.2
Length: 20 pages
Reserved DOI: 10.5281/zenodo.22728320 (record reserved)

Plain-language guide

3 minute explanation

What this paper is really saying.

The paper certifies quorum safety against correlated failures—such as one organization, region, or infrastructure provider failing—rather than pretending validators fail independently.

federated consensusfault domainscertification

Ten validators do not provide ten independent sources of resilience if eight share one cloud provider or legal operator. Ordinary validator counts can therefore exaggerate safety.

The proposed method declares failure domains and adversary policies explicitly, then searches for quorum pairs and domain covers that expose the smallest correlated failure capable of breaking intersection.

How it works

The argument, without the notation.

The paper’s technical details matter, but the basic route can be understood in three moves.

Catalog failure domains

Assign validators to organizations, regions, infrastructure providers, and other correlated-risk groups.

State the adversary policy

Define which combinations of domains may fail together and what counts as an admissible incident.

Search exact witnesses

Compute quorum pairs and minimum domain covers that certify or refute the desired safety margin.

What it adds

The useful takeaways.

Correlated failure is modeled directly instead of approximated through validator count.
The output includes exact witnesses for weak points and domain-cover margins for certified cases.
The supplied topology is an illustrative case study, not evidence about production Metriq infrastructure.

Scope & limits

What this does—and does not—establish.

Current status

Fault-domain certification methods and exact analysis of an illustrative topology. The case study is not production Metriq infrastructure evidence.

This report develops a conditional method and evaluates it within the stated evidence. It does not establish production performance, operational safety, or calibrated real-world predictive skill beyond that evidence.

A certificate is only as truthful as the declared failure-domain catalog.
The case study does not establish production-network resilience.
Unmodeled common causes can invalidate a seemingly strong margin.

Why it matters

Why anyone should care.

Counting validators can badly overstate resilience when many share one failure domain. Explicit domain-cover policies expose those correlated risks.

Key terms

The vocabulary, decoded.

These are the terms needed to understand the claim. The full paper uses them more precisely.

Fault domain

A group of components that may fail together because they share an operator, region, provider, or dependency.

Correlated failure

A failure event affecting several validators for one shared reason.

Adversary policy

The formal set of failure combinations a certification promises to tolerate.

Witness

A concrete quorum or fault configuration demonstrating that a claimed margin fails.

Independent review

Where scrutiny should concentrate.

The strongest review is not a general reaction. It tests the steps most capable of changing the conclusion.

01
Completeness and truth of the declared failure-domain catalog.
02
Domain-cover characterization and structured-fault assumptions.
03
Distinguish illustrative topology margins from production resilience.

Publication record

Go from explanation to evidence.

This page is a reading guide, not a substitute for the manuscript. The public record links the explanation to the paper, source package, review materials, and persistent identifier.

GitHub release Source folder

Document: Research Paper
Review status: Open for independent review
Published: 2026-09-12
DOI status: Reserved · 10.5281/zenodo.22728320
Canonical record: MF-PRISM-DLT-2026-04

Metriq PRISM Laboratory, Fault-Domain-Aware Federated Quorum Certification, Metriq PRISM Laboratory Research Paper MF-PRISM-DLT-2026-04, Version 0.2, 2026. Corresponding contributor: Daniel H. Jeffery, ORCID 0009-0001-1200-6042. DOI: 10.5281/zenodo.22728320.