S3 Gateway Performance Benchmark
Automated performance testing of the Fula decentralized S3-compatible storage gateway, compared against published industry baselines from AWS S3, IPFS, and Cloudflare R2.
About This Benchmark
The Fula S3 Gateway provides an S3-compatible API backed by decentralized IPFS storage. This benchmark measures real-world upload and download performance across different file sizes and concurrency levels, then compares the results against industry standards.
What is tested
- File sizes: 64 KB, 1 MB, and 10 MB objects
- Concurrency: 1 (sequential), 10, and 50 parallel requests
- Operations: Upload (PUT), Download (GET), and Mixed 80/20 read/write
- Metrics: Time to First Byte (TTFB), throughput (MB/s), latency percentiles, success rate
Industry baselines compared
- AWS S3 Standard — same-region EC2, published benchmarks
- IPFS Public Gateways — cached and uncached content retrieval
- Cloudflare R2 — S3-compatible API latency
How to Run the Benchmark
The benchmark is a self-contained Python script that tests any S3-compatible endpoint and generates a detailed markdown report with charts.
Prerequisites
pip install aiohttp matplotlib
# Python 3.9+ required
Run the benchmark
python benchmark_s3.py --token "YOUR_JWT_TOKEN" --endpoint https://s3.cloud.fx.land
Options
--token— (required) JWT Bearer token for authentication--endpoint— S3 endpoint URL (default:https://s3.cloud.fx.land)--bucket— Bucket name (default:benchmark-{timestamp})--output— Output report path (default:benchmark_results.md)
Output
The script produces a markdown report (benchmark_results.md) and
4 chart images (benchmark_*.png) in the same directory.
The report shown below is loaded directly from the latest generated results.