Used to do a couple of Search, but found no explanation of methods it really works stress tests.

So far as I realize, concurrency means synchronised quantity of connections.

ab most likely begins the given "c" quantity of connections at the identical time, and restrain the next waves until all demands are satisfied. Shall We Be Held right?

Given a greater quantity of demands, ab may perform demands/concurrency quantity of waves. Is correct?

Considering these claims, exactly what do these amounts mean:

  • Time/request
  • Time/request across all concurrent demands
  • Demands/second

(Essentially the 2nd measurement way to me that ab will be sending a brand new request in every concurrency level soon after the prior the first is satisfied.)

So I am confused. Thanks ahead of time.

I could not find almost anything to back this up, but I am confident that ab monitors concurrent demands and looks after a constant quantity of demands active, instead of delivering demands in waves. For instance, should you specify 50 concurrent connections, instead of awaiting the very first 50 demands to accomplish after which delivering another 50, ab could keep delivering a brand new request the moment a request completes, to ensure that there'll always be 50 active connections.