Published August 22, 2026 | Version v1

**Avalanche Confirmed\-State Latency Experiment: Reproducible Software, Protocol, Dataset and Analysis Pipeline**

  • 1. University of customs and finance

Description

This software project provides a complete reproducible framework for studying confirmed-state access latency in a permissioned Avalanche Layer-1 blockchain network. The project accompanies the research study “A Blockchain Solution for Reducing Confirmed-State Access Latency in an Avalanche Network: An Experimental Study” and implements the full workflow required to design, execute, analyze, verify, and reproduce the experiment.

The primary performance metric is T_visible, defined as the elapsed time between transmission of a signed blockchain transaction and the first successful retrieval of the accepted and updated state from an independent read node. Unlike conventional measurements based only on block interval or transaction receipt latency, this metric captures the complete operational path, including transaction submission, network propagation, block inclusion, consensus acceptance, execution, state commitment, and confirmed read-back.

The experimental protocol evaluates five block-production profiles across three network topologies and five offered-load levels using a randomized, paired design with repeated trials. The full registered campaign comprises 750 experimental runs and approximately 34.9 million scheduled transactions.

The repository contains the executable experimental protocol, workload generation tools, Avalanche network and subnet deployment configurations, measurement clients, smart-contract components, telemetry collection utilities, statistical analysis procedures, visualization modules, reproducibility tests, and deterministic Zenodo packaging tools. It also provides a documented simulated reference dataset that follows exactly the same schema as the intended field measurements.

A strict provenance mechanism distinguishes SIMULATED reference data from MEASURED experimental data. Simulated values are intended exclusively for validating the analytical workflow and must not be interpreted as measurements obtained from the physical cyber range. Once real transaction logs are available, they can be processed by the same analysis pipeline without changing the statistical methodology.

The project supports run-level statistical reduction, paired comparisons, bootstrap confidence intervals, multiple-comparison correction, pre-registered stability criteria, SHA-256 integrity verification, and deterministic regeneration of derived tables and figures. The software therefore provides both an experimental implementation and a reusable methodology for evaluating end-to-end confirmed-state visibility latency in permissioned blockchain infrastructures.

Keywords: Avalanche; blockchain; Layer 1; confirmed-state latency; transaction visibility; distributed systems; performance evaluation; cyber range; reproducible research; experimental software; blockchain benchmarking; latency measurement.

Files

avalanche-latency-experiment-code.zip

Files (1.6 MB)

Name Size
md5:965eb2f551bd8b5cc8c3c1b609fdc0f0
1.5 MB Preview Download
md5:1994d58f8bcd55d76a78ffc5b0ad1ad2
13.2 kB Preview Download