TS
Textsplain
explained in texts
Wednesday, May 27, 2026 · 9:41 AM
ok can you explain napkin math that catches bad designs like i actually need to build it
napkin math is how you notice that a cute architecture melts when traffic, storage, or fanout becomes real
mental picture: it is like checking if your moving truck can actually hold the apartment before you recruit six friends and vibes
good, now do the real thing
not the fake diagram version
estimate daily active users, actions per user, read/write ratio, peak multiplier, payload size, and retention window
convert daily events into requests per second so the system has a shape instead of a mood
so the first move is making the invisible numbers visible
exactly
storage math matters: a 2 KB object saved 100 million times is not a rounding error once indexes and replicas join
what changes when this gets real traffic?
bandwidth math catches chatty APIs, huge images, and designs that quietly send the same data thousands of times
latency budgets tell you which hops must be close, cached, parallelized, or moved out of the request path
wait, i thought the better answer was just adding the serious-sounding tool
that is the trap
the goal is not precise prediction. it is killing impossible designs early
ok so what bill shows up later?
overestimating wastes money and complexity
underestimating makes every later choice reactive
averages lie when traffic has celebrity spikes, launches, or time-zone waves
where do people usually mess this up?
using one big number like total users and forgetting peak traffic, object size, retention, and fanout
if i were designing this tomorrow, what should i write down first?
the read path
the write path
the thing that is allowed to be stale
the thing that absolutely is not
that is annoyingly practical
yeah. most system design is boring on purpose
write the assumptions down, round aggressively, and let the order of magnitude guide the architecture
got it
less architecture cosplay, more pressure map
perfect
draw the pressure, then choose the machinery
Read Wed, May 27 · 9:58 AM