napkin math

May 09, 2022

A newsletter about using napkin math and first-principle thinking to estimate systems performance — fast, and without writing any code! However, on a good day posts will include code to test whether the napkin math lines up with reality.

A table of performance ballpark numbers:

The archive page is a bit difficult to parse, as it has non-napkin-math posts scattered in it.

1: logging cost 2: database query latency 3: membership intersection service 4: redis throughput 5: composite primary keys 6: in-memory search 7: revision history 8: data synchronization 9: inverted index perf and merkle tree sync 10: mysql transactions per second vs fsyncs per second 11: circuit breakers 12: recommendations 13: filtering with inverted indexes 14: using checksums to verify syncing 100m database records 15: increase HTTP performance by fitting in the initial TCP window 16: when to write a simulator 17: neural network from scratch

↑ up