Traffic shaping is a technique used to control and prioritize network traffic, ensuring that bandwidth is allocated efficiently and fairly among different applications or users. In the Linux environment, traffic shaping plays a crucial role in optimizing network performance and providing a better user experience.

Linux offers various tools and mechanisms to implement traffic shaping, such as the Traffic Control (tc) command, which is part of the iproute2 package. With tc, you can define traffic classes, filters, and queuing disciplines to shape the network traffic according to your requirements.

To begin, you need to understand the key components of traffic shaping in Linux:

1. Traffic Classes: Traffic classes allow you to categorize network traffic based on specific criteria, such as protocol, source/destination IP address, or port number. By defining traffic classes, you can apply different shaping rules to each class, ensuring that critical traffic receives priority over non-critical traffic.

2. Filters: Filters are used to match packets based on specific criteria and classify them into appropriate traffic classes. Linux provides various filter options, such as u32, fw, and route, which allow you to match packets based on layer 3 (IP), layer 4 (TCP/UDP), or other header fields.

3. Queuing Disciplines: Queuing disciplines determine how packets are prioritized and scheduled for transmission. Linux offers a range of queuing disciplines, including pfifo, sfq, and htb, each with its own characteristics and algorithms. By configuring queuing disciplines, you can control the order in which packets are transmitted and ensure fair sharing of available bandwidth.

Now, let's look at some practical examples of traffic shaping in Linux:

Example 1: Limiting Bandwidth for a Specific Application Suppose you want to limit the bandwidth usage of a specific application, such as a file-sharing program. You can achieve this using the tc command as follows:

tc qdisc add dev eth0 root handle 1: htb default 10
tc class add dev eth0 parent 1: classid 1:1 htb rate 1mbit burst 15k
tc filter add dev eth0 parent 1: protocol ip prio 1 u32 match ip dport 6881 0xffff flowid 1:1

This example creates a traffic class (1:1) with a maximum rate of 1mbit and a burst size of 15k. The filter matches packets with a destination port of 6881 (commonly used by file-sharing programs) and assigns them to the traffic class.

Example 2: Prioritizing VoIP Traffic To prioritize VoIP traffic over other types of traffic, you can use the tc command with the prio queuing discipline. Here's an example:

tc qdisc add dev eth0 root handle 1: prio
tc filter add dev eth0 parent 1: protocol ip prio 1 u32 match ip dport 5060 0xffff flowid 1:1
tc filter add dev eth0 parent 1: protocol ip prio 2 u32 match ip dport 5060 0xffff flowid 1:2

In this example, the prio queuing discipline is applied to prioritize traffic. The first filter matches packets with a destination port of 5060 (commonly used by VoIP) and assigns them to flowid 1:1. The second filter matches packets with a destination port of 5060 and assigns them to flowid 1:2. This ensures that VoIP traffic is given higher priority.