Beyond Software Optimizations: Breaking 100us Barrier in EC2
Details
Getting to 100µs latencies takes more than just software optimization. It takes deep understanding of the kernel, scheduling, and the networking stack. In this talk, we cover our journey on getting sub 100µs latencies inside EC2. We will cover EC2 placement groups, core pinning, IRQs, and EC2 interrupt moderation. You will learn the knobs that you can tune and their impact with real world data.
Food and drinks provided!
