Cold-Start Saving

Problem

Compute the latency saved by caching a warm model: expected = cold * (1 - hit_rate) + warm * hit_rate, and saving = cold - expected.

The signature is cold_start_saving(warm_ms, cold_ms, cache_hit_rate).

Hint

Keeping replicas warm is one of the most effective cold-start mitigations.

case 1

in: [100,1000,0.5]

expected: 450

case 2

in: [100,1000,0]

expected: 0

case 3

in: [100,1000,1]

expected: 900

case 4

in: [200,500,0.8]

expected: 240

Starter codePython

def cold_start_saving(warm_ms, cold_ms, cache_hit_rate):
    # Your code here
    pass

4 test cases · run your code against them in the editor