most rpc benchmarks are horribly flawed
they are not global, run over home internet, ignore tail latencies, omit data freshness/correctness, and cherry-pick time-ranges
today we're releasing a new, open-source RPC benchmarking tool
featuring:
- 24/7 coverage
- randomized inputs (cannot be cached/gamed)
- freshness & correctness checks
- multi-provider support (TSW, AWS, GCP, CF)
- global coverage (US, EU, APAC)
- user-specific scoring method (configurable)