hft - tuning guide

@ kernel level

Tuning for low latency systems is a multi-layer approach. Attention is paid to different components spanning hardware, firmware, OS/kernel, network stack, application layer logic, toolchain, network infra investments, memory management etc.

On a tangent, other techniques/agreements such as micro benchmarking, hardware timestamping, mm contracts come to mind as important. On a rather precariously long list of items, I want to break them down by layer.

My philosophy around this research/blog is if it’s useful to me, it’s useful for someone else. I wanted a nice compilation as a personal handbook. Think it will be pretty useful for a lot of readers, hence I will compile them in the market notes down the line.

This post is dedicated to the (linux) kernel component.