Your graph showing the RTT against the 100ms target doesn't account for the other costs of getting a page in front of someone. This would be more convincing if you had a latency budget that was already partly spent. You can use performance metrics on the current page load to seed that value, noting that this page is already very fast by virtue of it being static.
Your graph showing the RTT against the 100ms target doesn't account for the other costs of getting a page in front of someone. This would be more convincing if you had a latency budget that was already partly spent. You can use performance metrics on the current page load to seed that value, noting that this page is already very fast by virtue of it being static.