There is a server on AWS that I do not control that I need as low latency to as possible. I figured I would experimentally measure my latency from different AWS regions to narrow down the lowest latency region for my instance. I have reasons to suspect that there is still large latency gains that can be made.
A traceroute shows that the traffic is going through an Amazon Global Accelerator (AGA), so my current thought it that lowest possible latency would be achieved by collocating with AGA.
I was wondering if there is a better known approach for optimizing latency to applications behind AGA? Would activating AGA on my own instance give better latency to other instances behind AGA, or does it not do anything for outbound traffic?
