Chapter 6. Interfaces and Subnets

Table of Contents

6.1. Relationship between Interfaces and Physical Ports
6.1.1. Port groups
6.1.2. Interfaces
6.2. Defining port groups
6.3. Defining an interface
6.3.1. Defining subnets
6.3.1.1. Source filtering
6.3.1.2. Using DHCP to configure a subnet
6.3.1.3. Using SLAAC (IPv6 router announcements) to configure a subnet
6.3.1.4. Providing IPv6 addresses to devices on a network (IPv6 router announcements)
6.3.1.5. IPv6 prefix delegated subnets
6.3.2. Setting up DHCP server parameters
6.3.2.1. Fixed/Static DHCP allocations
6.3.2.2. Restricted allocations
6.3.2.3. Special DHCP options
6.3.2.4. Logging
6.3.3. DHCP Relay Agent
6.4. Physical port settings
6.4.1. Disabling auto-negotiation
6.4.2. Setting port speed
6.4.3. Setting duplex mode
6.4.4. Defining port LED functions

This chapter covers how to set up Ethernet interfaces and the definition of subnets that are present on those interfaces.

For information about other types of 'interfaces', refer to the following chapters :-

6.1. Relationship between Interfaces and Physical Ports

The FB2700 features four Gigabit Ethernet (1Gb/s) ports that can also operate at 10Mb/s and 100Mb/s speeds. Auto-negotiation of link speed is enabled by default, so when connected to auto-negotation capable equipment, the ports operate at the highest speed that both ends of the link can run at. In some situations, auto-negotiation is not supported by connected equipment, and so the FB2700 provides control of port behaviour to allow the port to work with such equipment.

Each port features a green and amber LED, the functions of which can be chosen from a range of options indicating link speed and/or traffic activity.

The exact function of the ports is flexible, and controlled by the configuration of the FB2700.

6.1.1. Port groups

Up to four port groups can be defined, with each group comprising a set of one or more physical ports that doesn't overlap with any other group. The ports within the group normally work as a conventional Ethernet switch, directly transferring traffic at wire-speed that is destined for a MAC address that is present on one of the other ports in the group, but can alternatively be configured as a trunked group.

Note

The same MAC responding to ARP requests on multiple physical ports (even with different VLANs) within a port group is not supported.

The port group has a trunk setting which defaults to being false. When only one port is in the group it makes no difference how this is set. With more than one port, when trunk is false, the ports work as a switch, passing traffic directly at gigabit speeds between the physical ports. With more than one port, when trunk is true, the ports work as a link aggregation trunk and not as a switch. There is no option for some ports in a group to be trunked and also switched to other ports.

The FireBrick supports LACP (Link Aggregation Control Portocol) which is used to coordinate and control trunked port groups by exchanging LACP packets over the links. There is a lacp setting in the individual ethernet port settings which can be used to control LACP's behaviour, as follows:

  • lacp="false": It is assumed that the link is not connected to a device supporting LACP. LACP packets are not sent, and any received are ignored. The ports in a trunked port group will be used for aggregation when the physical link is up, after a short delay to ensure the partner is ready.
  • lacp="true": The link must be connected to an LACP-enabled device in order to function. LACP packets are sent, and the link will only be enabled for traffic when LACP negotiation is successful.
  • lacp not set: This is the default (Auto) setting. LACP packets will be sent if the port is part of a trunked port group, or if LACP packets are detected from the linked device. If LACP is not detected, a non-trunked port will always be enabled, while a port which is part of a trunked port group will only be enabled if it is the lowest-numbered (leftmost) port in the group. There will be a short delay after the port is physically up to allow for detection of LACP. When LACP is detected, the LACP negotiation controls the availability of the port.

The FB2700 has an internal 1G switch with a single link to communicate with the CPU. The maximum bandwidth for incoming traffic is limited by that link, and so dedicating multiple physical ports in a bonded way to LACP could result in the switch receiving twice as much traffic as it can send onwards. This would result in the internal switch dropping packets during traffic spikes. To avoid this, we configure any additional ports as a hot standby. This provides rapid failover whilst allowing the other device a chance to shape/queue packets appropriately.

6.1.2. Interfaces

In the FB2700, an interface is a logical equivalent of a physical Ethernet interface adapter. Each interface normally exists in a distinct broadcast domain, and is associated with at most one port group.

Each port group, which could be a single port, can operate simply as an interface with no VLANs, or can have one or more tagged VLANs which are treated as separate logical interfaces. Using VLAN tags and a VLAN capable switch you can effectively increase the number of physical ports.

Appendix E contains a brief overview of VLANs and the concept of broadcast domains.

By combining the FB2700 with a VLAN capable switch, using only a single physical connection between the switch and the FB2700, you can effectively expand the number of distinct physical interfaces, with the upper limit on number being determined by switch capabilities, or by inherent IEEE 802.1Q VLAN or FB2700 MAC address block size. An example of such a configuration is a multi-tenant serviced-office environment, where the FB2700 acts as an Internet access router for a number of tenants, firewalling between tenant networks, and maybe providing access to shared resources such as printers.