Skip to content

README.md

Latest commit

 

History

History
54 lines (39 loc) · 1.82 KB

README.md

File metadata and controls

54 lines (39 loc) · 1.82 KB
Raw

Benchmark Data

This directory contains compact raw artifacts suitable for version control and manifests describing the larger JMeter datasets stored outside Git.

Public raw data

phase2-iperf/ contains the complete iperf3 JSON output for the canonical three-run TCP congestion-control campaign.

The files use descriptive names:

  • cubic-clean-runN.json
  • cubic-loss-1pct-runN.json
  • bbr-clean-runN.json
  • bbr-loss-1pct-runN.json
  • cubic-delay50ms-loss2pct-runN.json
  • bbr-delay50ms-loss2pct-runN.json

Receiver throughput is used as the primary TCP throughput metric.

External raw data

The JMeter JTL files and execution logs are stored outside Git because the complete local dataset is several gigabytes.

The canonical external campaigns are:

  • Phase 1: three runs for bare metal, Docker host networking, and Docker bridge networking.
  • Phase 3: five runs for each equal-CPU architecture at 300, 480, and 840 requests per second.
  • Benchmark-validity case study: the invalidated JMeter throughput-ceiling campaign.

Execution logs remain preserved in the external archive. They are not included in the public repository because the per-run CSV files, aggregate summaries, configuration files, checksums, and analysis code provide a smaller and more reviewable evidence set.

Manifests

  • manifests/canonical-raw-artifacts.sha256 records SHA-256 checksums.
  • manifests/canonical-raw-artifacts.tsv records checksum, file size, path, and storage classification.
  • manifests/phase2-public-json.sha256 records checksums for committed iperf3 JSON files.

Processed per-run and aggregate outputs are published in results/processed/.

A public clone can validate processed outputs and regenerate figures. Full Phase 1, Phase 3, and benchmark-validity result regeneration requires access to the external JMeter dataset.