The email sent will contain a link to this article, the article title, and an article excerpt (if available). For security reasons, your IP address will also be included in the sent email.
How do you systematically identify the sources of jitter and remove them one by one in a low-latency trading system? That was the question asked on the mechanical-sympathy email list.
Gil Tene, Vice President of Technology and CTO, Co-Founder, Azul Systems gave the sort of answer that comes from the accumulated wisdom born from lots of real-life experience. It's an answer that needed sharing. And here it is:
Finding the cause of hiccups/jitters in a a Linux system is black magic. You often look at the spikes and imagine "what could be causing this".
Based on empirical evidence (across many tens of sites thus far) and note-comparing with others, I use a list of "usual suspects" that I blame whenever they are not set to my liking and system-level hiccups are detected. Getting these settings right from the start often saves a bunch of playing around (and no, there is no "priority" to this - you should set them all right before looking for more advice...).
My current starting point for Linux systems that are interested in avoiding many-msec hiccup levels is:
Return to Article