6. SR Linux interfaces

This chapter describes SR Linux interface types, subinterfaces, and support for Link Aggregation Groups (LAGs). See the “Interfaces” chapter in the SR Linux Configuration Basics Guide for configuration examples.

6.1. SR Linux interface types

On the SR Linux, an interface is any physical or logical port through which packets can be sent to or received from other devices.

The SR Linux supports the following interface types:

  1. Loopback
    A loopback interface is a virtual interface that is always up, providing a stable source or destination from which packets can always be originated or received. The SR Linux supports up to 256 loopback interfaces system-wide, across all network-instances. Loopback interfaces are named loN, where N is 0 to 255.
  2. Network
    Network interfaces carry transit traffic, as well as originate and terminate control plane traffic and in-band management traffic.
    The physical ports in line cards installed in the SR Linux are network interfaces. A typical line card has a number of front-panel cages, each accepting a pluggable transceiver. Each transceiver may support a single channel or multiple channels, supporting one Ethernet port or multiple Ethernet ports, depending on the transceiver type and its breakout options.
    In the SR Linux CLI, each network interface has a name that indicates its type and its location in the chassis. The location is specified with a combination of slot number and port number, using the following formats:
    ethernet-slot/port
    For example, interface ethernet-2/1 refers to the line card in slot 2 of the SR Linux chassis, and port 1 on that line card.
  3. Management
    Management interfaces are used for out-of-band management traffic. The SR Linux supports a single management interface named mgmt0.
    The mgmt0 interface supports the same functionality and defaults as a network interface, except for the following:
    1. Packets sent and received on the mgmt0 interface are processed completely in software.
    2. The mgmt0 interface does not support multiple output queues, so there is no output traffic differentiation based on forwarding class.
    3. The mgmt0 interface does not support pluggable optics. It is a fixed 10/100/1000-BaseT copper port.
  4. Integrated routing and bridging (IRB)
    IRB interfaces enable inter-subnet forwarding. Network instances of type mac-vrf are associated with a network instance of type ip-vrf via an IRB interface.

6.2. LAG interfaces

A Link Aggregation Group (LAG), based on the IEEE 802.1ax standard (formerly 802.3ad), increases the bandwidth available between two network devices, depending on the number of links installed. A LAG also provides redundancy in the event that one or more links participating in the LAG fail. All physical links in a given LAG links combine to form one logical interface.

LAGs can be either statically configured, or formed dynamically with Link Aggregation Control Protocol (LACP). Load sharing is executed in hardware, which provides line rate forwarding for all port types. A LAG consists of ports of the same speed.

6.3. Subinterfaces

On the SR Linux, each type of interface can be subdivided into one or more subinterfaces. A subinterface is a logical channel within its parent interface.

Traffic belonging to one subinterface can be distinguished from traffic belonging to other subinterfaces of the same port using encapsulation methods such as 802.1Q VLAN tags.

While each port can be considered a shared resource of the router that is usable by all network-instances, a subinterface can only be associated with one network-instance at a time. To move a subinterface from one network-instance to another, you must disassociate it from the first network-instance before associating it with the second network-instance.

You can configure ACL policies to filter IPv4 and/or IPv6 packets entering or leaving a subinterface.

The SR Linux supports policies for assigning traffic on a subinterface to forwarding classes or remarking traffic at egress before it leaves the router. DSCP classifier policies map incoming packets to the appropriate forwarding classes, and DSCP rewrite-rule policies mark outgoing packets with an appropriate DSCP value based on the forwarding class.