Understanding BGP Route Selection
Posted on July 3, 2026 · 8 min read
Border Gateway Protocol (BGP) is the glue that holds the Internet together. When multiple paths exist to the same destination, BGP uses a well-defined tie-breaking algorithm to select the best route. Understanding this process is essential for any network engineer working with multi-homed environments.
The selection process starts with the highest local-preference, followed by the shortest AS_PATH, then the lowest origin type. Only after these are the MED and IGP metrics considered.
networkingbgprouting
Kernel Tuning for High-Throughput Proxies
Posted on June 28, 2026 · 12 min read
When operating a high-throughput TCP proxy, default kernel parameters often become a bottleneck. Adjusting tcp_rmem, tcp_wmem, and the backlog queues can dramatically improve throughput under load.
Key parameters to review include net.core.somaxconn, net.ipv4.tcp_max_syn_backlog, and the file descriptor limit via ulimit -n. These small changes can mean the difference between serving 10k and 100k concurrent connections.
linuxperformancetuning
DNS-over-HTTPS Deployment Patterns
Posted on June 20, 2026 · 6 min read
DNS-over-HTTPS (DoH) is becoming the standard for private DNS resolution. Setting up a local DoH proxy with caching and upstream fallback provides both privacy and resilience against DNS-based blocking or manipulation.
dnsprivacyinfrastructure