Benchmarking with:

Which option is best overall? Cross-engine rankings.
The options are sorted top-down by performance/cost ratio, most efficient at the top of the list.
We conduct a standard HammerDB TPC-C benchmark.
The disk benchmarks are done with fio arguments --name=readwrite --rw=randrw --size=512M --numjobs=4 --time_based --runtime=60 --group_reporting.
- where configurable we use a default size of 1000 GByte SSD for storage
- for benchmarking we calculate the number of vUsers by vcores x 2
- for benchmarking we calculate the number of Warehouses by vUsers x 5
Limited data preview
Part of the benchmark results are hidden. To fully unlock all data, apply for a subscription.
| Provider | DB Engine | Deployment | vCPU / RAM | Availability | NewOrders/minute | Transactions/minute | Monthly cost / 1000 tpm | Benchmark |
|---|
