109.3 Lesson 1

The ip command is a fairly recent utility used to view and configure just about anything relating to network configurations. This lesson covers some of the most used subcommands of ip, but it barely scratches the surface of what is available. Learning to read the documentation will help you be much more efficient with it.

Each subcommand of ip has its own man page. The SEE ALSO section of the ip man page has a list of them:

Instead of looking at this every time you need the man page, simply add - and the name of the subcommand to ip, e.g. man ip-route.

Another source of information is the help function. To view the built-in help, add help after the subcommand:

IPv4 and IPv6 are what are known as routed or routable protocols. This means they are designed in a way that make it possible for network designers to control traffic flow. Ethernet is not a routable protocol. This means that if you were to connect a bunch of devices together using nothing but Ethernet, there is very little you can do to control the flow of network traffic. Any measures to control traffic would end up similar to current routable and routing protocols.

Routable protocols allow network designers to segment networks to reduce the processing requirements of connectivity devices, provide redundancy, and manage traffic.

IPv4 and IPv6 addresses have two sections. The first set of bits make up the network section while the second set make up the host portion. The number of bits that make up the network portion are determined by the netmask (also called subnet mask). Sometimes it will also be referred to as the prefix length. Regardless of what it is called, it is the number of bits that the machine treats as the network portion of the address. With IPv4, sometimes this is specified in dotted decimal notation.

Below is an example using IPv4. Notice how the binary digits maintain their place value in the octets even when it is divided by the netmask.

The network portion of an address is used by an IPv4 or IPv6 machines to lookup which interface a packet should be sent out on in its routing table. When an IPv4 or IPv6 host with routing enabled receives a packet that is not for the host itself, it attempts to match the network portion of the destination to a network in the routing table. If a matching entry is found, it sends the packet to the destination specified in the routing table. If no entries are found and a default route is configured, it is sent to the default route. If no entry is found and no default route are configured, the packet is discarded.

There are two tools we will be covering that you can use to configure a network interface: ifconfig and ip. The ifconfig program, while still widely used, is considered a legacy tool and may not be available on newer systems.

Before configuring an interface, you must first know what interfaces are available. There are a few ways to do this. One way is to use the -a option of ifconfig:

Another way is with ip. Sometimes you will see examples with ip addr, ip a, and some with ip address. They are synonymous. Officially, the subcommand is ip address. This means that if you wish to view the man page, you must use man ip-address and not man ip-addr.

The link subcommand for ip will list the interface links available for configuration:

Assuming the sys filesystem is mounted, you can also list the contents of /sys/class/net:

To configure an interface with ifconfig, you must be logged in as root or use a utility such as sudo to run the command with root privilege. Follow the example below:

The Linux version of ifconfig is flexible with how you specify the subnet mask:

Notice how with IPv6 the keyword add was used. If you don’t precede an IPv6 address with add, you will get an error message.

The following command configures an interface with ip:

With ip, the same command is used for both IPv4 and IPv6.

The commands route, netstat -r, and ip route can all be used to view your routing table. If you wish to modify your routes, you need to use route or ip route. Below are examples of viewing a routing table:

Notice how there is no output regarding IPv6. If you wish to view your routing table for IPv6, you must use route -6, netstat -6r, and ip -6 route.

An example of netstat -r6 has been omitted because its output is identical to route -6. Some of the output of the above route command is self explanatory. The Flag column provides some information about the route. The U flag indicates that a route is up. A ! means reject route i.e. a route with a ! won’t be used. The n flag means the route hasn’t been cached. The kernel maintains a cache of routes for faster lookups separately from all known routes. The G flag indicates a gateway. The Metric or Met column isn’t used by the kernel. It refers to the administrative distance to the target. This administrative distance is used by routing protocols to determine dynamic routes. The Ref column is the reference count, or number of uses of a route. Like Metric, it is not used by the Linux kernel. The Use column shows the number of lookups for a route.

In the output of netstat -r, MSS indicates the maximum segment size for TCP connections over that route. The Window column shows you the defualt TCP window size. The irtt shows the round trip time for packets on this route.

The output of ip route and ip -6 route reads as follows:

Working through a few examples should clarify this: