Extreme HTTP Performance Tuning: 1.2M API req/s on a 4 VCPU EC2 Instance

• raids

  2 4 6.1 C++

  Cache distribution services: http, websock, redis, memcached

  

Great work, thanks!

I'm curious whether disabling the slow kernel network features competes with an tcp bypass stack. I did my own wrk benchmark [0], but I did not try to optimize the kernel stack beyond pinning CPUs and busypoll, because the bypass was about 6 times as fast. I assumed that there is no way the kernel stack could compete with that. This article shows that I may be wrong. I will definitely check out SO_ATTACH_REUSEPORT_CBPF in the future.

[0] https://github.com/raitechnology/raids/#using-wrk-httpd-load...

• wrk

  36 36,760 0.0 C

  Modern HTTP benchmarking tool

pMAX 626.00us

If you were to apply wrk's coordinated omission hack to these result, the backfilling only starts for requests that took longer than p50 x 2 (roughly) = 406us, which is probably somewhere between p99.999 and pMAX; a very, very small percentage.

I am not claiming that wrk's hack is "correct", just that I don't think coordinated omission is a major concern for *this specific workload/environment*

1. https://github.com/wg/wrk/blob/a211dd5a7050b1f9e8a9870b95513...

• InfluxDB

  www.influxdata.com featured

  Power Real-Time Data Analytics at Scale. Get real-time insights from all types of time series data with InfluxDB. Ingest, query, and analyze billions of data points in real-time with unbounded cardinality.

  

Suggest a related project