3byte

cpu-bound single-threaded app server on server with 16 logical cpu cores. -> fork() every req? fixed pool of server procs?
when does read-write lock outperform simple mutex? -> write-heavy load, read-heavy load, ...
how to increase write load on sql database? -> horizontal scaling, indices, ...
why do programs with bad memory leaks slow down before they crash? -> swapping, fragmentation, ...
for i: 0-9, settimeout(...), 1000: what does this do?
how do long-polling and websockets compare?
store the locations of 20k users with high read/write throughput, don't care about loss. -> redis/memcached over fs/db/py