Performance results - Private vs. Public

WebPageTest Private Instances
1

Why are my Performance results so different?

PRIVATE
Performance Results (Median Run)
Document Complete Fully Loaded
Load Time First Byte Start Render User Time Speed Index DOM Elements Time Requests Bytes In Time Requests Bytes In Certificates
First View (Run 2) 8.325s 0.539s 1.190s 8.414s 2419 2605 8.325s 241 3,355 KB 11.561s 252 3,391 KB 81 KB
Repeat View (Run 1) 4.425s 0.456s 0.789s 4.515s 2165 2609 4.425s 86 954 KB 6.890s 89 2,419 KB 39 KB

PUBLIC
Document Complete Fully Loaded
Load Time First Byte Start Render Speed Index Time Requests Bytes In Time Requests Bytes In Certificates Cost
First View 23.893s 1.342s 13.476s 15878 23.893s 228 2,906 KB 43.485s 235 2,939 KB 94 KB $$$$$
Repeat View 17.573s 0.603s 2.282s 5685 17.573s 78 582 KB 29.374s 84 583 KB 41 KB

2

How many runs did you do for each, were they run at the same time and from similar locations? The 17-23s first byte time indicates a server response time problem on the “public” tests that was still pretty bad but not as bad when you ran the private tests.

3

Sorry, looked at the wrong column :frowning:

4

What kind of machine/vm were the private tests run on and what was the public test run on (and from what locations)?

The network connectivity (routing and distance to the server) and CPU speed on the test machines can both radically impact the performance.

5

Private instance is in local work network running over a mac mini, and for public I used Phoenix, AZ location, both in Chrome, a single run for each.

6

I’d recommend running 9 runs and see what the stability of the results looks like. Also, if you use “Dulles Thinkpad” that is probably closest hardware-wise to the Mac Mini.