Anatomy of two Linux I/O models. epoll vs io_uring, readiness vs completion, syscall overhead, SQPOLL — and when reaching for io_uring instead of an event loop actually pays off.
epoll vs io_uring — When the Event Loop Isn’t Enough Read Post »
Skip the container abstraction layer. Enforce CPU, memory & I/O limits directly via cgroupfs — zero daemon overhead, no OCI ceremony. A deep dive into cgroups v2 on Arch Linux.
cgroups v2 as a Native Runtime Isolation Primitive — No Docker Required Read Post »
fork() in Linux isn’t just process copying. It’s one of the cleverest kernel mechanisms — Copy-on-Write, shared file descriptors, inherited mutexes. Find out what really happens under the hood.
Linux fork() — how it works, copy-on-write and zombie processes Read Post »