
About the importance of IT caches
In the discussion, the problem was spontaneously born, it seemed interesting:
A huge service performs requests, the results of which are cached. Executing an un cached request is tens of thousands of times longer than a cached request. Once, something went bad and the percentage of hits in this cache fell from 99% to 98%.
Previously, the service worked on 100 absolutely identical servers. How many servers will be required now?
PS: the task is not quite strictly formulated, but this should not greatly interfere with its solution :)
A huge service performs requests, the results of which are cached. Executing an un cached request is tens of thousands of times longer than a cached request. Once, something went bad and the percentage of hits in this cache fell from 99% to 98%.
Previously, the service worked on 100 absolutely identical servers. How many servers will be required now?
PS: the task is not quite strictly formulated, but this should not greatly interfere with its solution :)