Prior to this, I had already thought a bit about measuring response times against a set goal as described in issue #60. (I created that issue just now after this and a couple of other issues.)
Arbitrarily chosen values:
16 ms for requests which do not write timelog data.
32 ms for requests which do write timelog data.
To be measured by tests (see also issue #35) connecting through Nginx. Time is from request is sent until response is received. Note that the budget does not make any claims about timing for remote clients. What ever extra time packets take to travel over the network is beyond the scope of this budget.
The idea of explicit resource budgets was brought up in the March 2010 Addison-Wesley Professional book The Design of Design: Essays from a Computer Scientist by Frederick P. Brooks Jr. I recently bought and am currently reading in eBook (ePub) format. http://www.informit.com/store/design-of-design-essays-from-a-computer-scientist-9780201362985
Prior to this, I had already thought a bit about measuring response times against a set goal as described in issue #60. (I created that issue just now after this and a couple of other issues.)
Arbitrarily chosen values:
To be measured by tests (see also issue #35) connecting through Nginx. Time is from request is sent until response is received. Note that the budget does not make any claims about timing for remote clients. What ever extra time packets take to travel over the network is beyond the scope of this budget.