Optimizing Linux Network Performance

Introduction

Network performance is crucial for many Linux systems, whether they're used as servers, workstations, or embedded devices. This guide will walk you through various techniques and tools to optimize your Linux system's network performance.

Understanding Network Performance Factors

Network Interface Card (NIC) capabilities
TCP/IP stack configuration
Kernel network parameters
Network drivers and firmware
System resource allocation
Network protocols and services in use

Basic Network Optimization Techniques

1. Update Network Drivers

Ensure you have the latest network drivers for your hardware:

sudo lshw -C network

Check for and install any available updates for your network hardware.

2. Enable TCP BBR (Bottleneck Bandwidth and Round-trip propagation time)

BBR can significantly improve network throughput and latency:

echo "net.core.default_qdisc=fq" | sudo tee -a /etc/sysctl.conf
echo "net.ipv4.tcp_congestion_control=bbr" | sudo tee -a /etc/sysctl.conf
sudo sysctl -p

3. Adjust Network Interface Queue Length

Increase the queue length for better performance under high load:

sudo ip link set dev eth0 txqueuelen 10000

4. Optimize Network Buffers

Adjust the following sysctl parameters:

net.core.rmem_max = 16777216
net.core.wmem_max = 16777216
net.ipv4.tcp_rmem = 4096 87380 16777216
net.ipv4.tcp_wmem = 4096 65536 16777216

Advanced Network Optimization

1. Network Interface Card (NIC) Tuning

Enable/disable offloading features
Adjust interrupt coalescence
Configure Receive Side Scaling (RSS)

Example (adjust based on your NIC capabilities):

sudo ethtool -K eth0 tso on gso on gro on

2. IRQ Balancing

Distribute network interrupts across multiple CPU cores:

sudo apt install irqbalance
sudo systemctl enable --now irqbalance

3. Use TCP Fast Open

Enable TCP Fast Open for faster connection establishment:

echo 3 | sudo tee /proc/sys/net/ipv4/tcp_fastopen

4. Optimize for High-Speed Networks

For 10Gbps+ networks, consider increasing the following:

net.core.netdev_max_backlog = 30000
net.core.somaxconn = 1024

Tip:

Always test thoroughly after making changes. Network optimizations can have different effects depending on your specific hardware, workload, and network environment.

Monitoring and Benchmarking

Use these tools to monitor and benchmark your network performance:

Tool	Purpose	Example Command
iperf3	Measure network throughput	`iperf3 -c server_ip`
nethogs	Monitor per-process network usage	`sudo nethogs`
nload	Monitor network traffic and bandwidth usage	`nload`
tcpdump	Analyze network traffic	`sudo tcpdump -i eth0`

Security Considerations

While optimizing network performance, keep these security aspects in mind:

Regularly update your system and network drivers
Use a firewall (e.g., iptables or nftables)
Implement proper access controls and authentication
Consider using VPNs for sensitive traffic
Monitor your network for unusual activity

Note:

Network optimization is an ongoing process. Regularly review and adjust your settings as your network requirements and infrastructure change.

Troubleshooting Common Issues

High latency: Check for network congestion, DNS issues, or routing problems
Packet loss: Investigate physical connections, driver issues, or network congestion
Low throughput: Verify NIC settings, TCP window sizes, and potential bottlenecks
Connection drops: Check for hardware issues, driver problems, or misconfigured network parameters

Additional Resources

Scroll to Top