Cloud Runtime Bench
A TypeScript benchmarking harness that deploys the same JavaScript backend to Docker/Node, Oracle Cloud Functions, Vercel Functions and WasmEdge and measures latency, throughput and cold starts across CPU, memory, IO and API workloads, with Pulumi and docker-compose infrastructure.
- BPM
- 139
- GRADE
- B
- LEVEL
- 5
- RELEASES
- 0
- YEARS
- 2025–2025
Same Code, Four Clouds
In May 2025 I wanted real numbers on cold starts, so I built a harness: one JavaScript backend, deployed unchanged to Docker on Node, Oracle Cloud Functions, Vercel Functions and WasmEdge, then hit with the same four scenarios.
CPU work, memory pressure, IO, and a plain API call, measuring latency, throughput, and how long each platform takes to wake up from nothing. Pulumi for the cloud side and docker-compose for the local side, so every run was reproducible. All of it TypeScript.
WasmEdge is in there because I was thinking about where JavaScript could run at all, which was on my mind that month between the egui bridges and the start of tumeric. A year later Tish compiled to native code, JavaScript, and WebAssembly, and this was me checking out those destinations first.