HTTP2's effect on performance
A recent upgrade to my hosting enabled HTTP2 so I took the opportunity to see how this change affected performance of my site.
As part of routine monitoring I run tests on my homepage which is very well optimised for performance - to the point where time to first byte is the limiting factor. As a result a small change to this will show up, especially given the time dilation effect produced by a connection throttled to 3G.
As with other subjective timing measures there's a lot of variation and the differences can be well within the range of variation so these times are the median of three runs on each day and I'm comparing over 10 day periods each side of the change.
| Day | First Byte | |
|---|---|---|
| HTTP2 | HTTP1.1 | |
| Average | 1.87 | 2.23 |
| 1 | 2.13 | 2.13 |
| 2 | 1.89 | 2.49 |
| 3 | 1.83 | 2.15 |
| 4 | 2.10 | 1.88 |
| 5 | 1.59 | 1.87 |
| 6 | 2.06 | 1.89 |
| 7 | 1.81 | 2.45 |
| 8 | ---- | 2.20 |
| 9 | 1.84 | 2.46 |
| 10 | 1.62 | 2.74 |
That gives a difference between the averages showing that HTTP2 is 0.35s faster. Whilst that's short of a statistically significant result - mainly because there's not many data points in each set - I'm happy enough that it's validated the change and I can move on to testing the next change.
Want some more help with optimising your web performance? Get in touch.