The Linux network stack tops out at ~2 Mpps per core. DPDK reaches 20 Mpps. The difference isn't speed — it's removing work: no syscalls, no allocations, no demux, no locks. Why kernel-bypass exists and when it earns its complexity.
When recv() returns, dozens of layers ran behind your back — DMA, NAPI, sk_buff, protocol demux, the socket queue. A trace of the entire receive path with the cost of each hop and where the latency actually lives.
select and poll were O(n). epoll is O(1). The reason isn't smarter scanning — it's that the kernel pushes ready events into a list as they happen. A walk through the data structures that make it work.
Using an ESP32 and ESP-IDF to build a network-aware Wake-on-LAN trigger — power on machines remotely over MQTT without an always-on server in between.
A deep dive into implementing a functional TCP/IP stack entirely in software — covering Ethernet framing, ARP, IP routing, TCP state machines, and the painful lessons learned along the way.