Performance Tuning & Kernel Parameters (Sysctl)

Stop your network throughput from hitting a wall with BBR

🧩 The Challenge

Dealing with high-latency links where your transfers crawl even though you have plenty of bandwidth is infuriating. I spent a whole afternoon debugging a supposedly fast connection that was dying under packet loss until I remembered TCP congestion control algorithms.

💡 The Fix

Switching your congestion control algorithm to BBR helps your system handle packet loss and high latency much more gracefully. It is essentially free speed for most high-traffic server workloads.

sysctl -w net.core.default_qdisc=fq
sysctl -w net.ipv4.tcp_congestion_control=bbr

⚙️ Why It Works

By switching to BBR, the kernel stops assuming all packet loss is congestion and instead models the bottleneck bandwidth and round-trip time. You’ll find it pushes data through congested paths way faster than the default cubic algorithm ever could.

🚀 Pro-Tip: Don’t forget to add these to /etc/sysctl.conf so your settings don’t vanish into thin air after a reboot.

Linux Tips & Tricks | © ngelinux.com | 8/28/2026