Day 4 results: the queue and the cliff
Measured on: (machine, OS, Python version)
Output
Paste part 1, part 2 and the scoreboard here.
The curve, drawn by hand
Utilization on the x axis, average latency on the y axis. Photo here.
One sentence per prediction
- Q1 (50% busy):
- Q2 (90% busy):
- Q3 (99% busy):
- Q4 (two servers):
Little’s Law check
Counted in system: __ Little’s Law said: __ How close, and why:
Part 2 vs part 1
Where did the real thread-based numbers drift from the simulation, and why: