VRF Support in DHCPv4

Virtual Routing and Forwarding (VRF) controls information flow in a network by partitioning the network into different logical VRF domains to isolate traffic.

One device can have multiple containers of routing tables or Forwarding Information Bases (FIBs), with one routing table for each VRF instance. A VRF-capable router can function as a group of multiple virtual routers on the same physical router.

Inter-VRF route leaking allows leaking of specific route prefixes from one VRF instance to another on the same physical router, which eliminates the need for external routing.

In a DHCP setting, route leaking is controlled through one DHCP server (which may be on a different VRF) so that multiple VRFs can communicate with that server.

DHCP relay is supported in the following configurations.

Same VRF instance as the interface through which the client is connected
For example:
  • VE interface 100 in VRF "red"
  • IP address of interface - 3.1.1.1/24
  • IP DHCP Relay address (20.1.1.2)
Different VRF instance as the interface through which the client is connected, an inter-VRF deployment
For example:
  • VE interface 100 in default VRF
  • IP address of interface - 3.1.1.1/24
  • IP DHCP Relay address (100.1.1.2) in VRF "blue"
  • IP DHCP Relay address (1.2.3.4.6) in VRF "red"
A maximum of 128 inter-VRF IP DHCP Relay address configurations is allowed per node. A VRF route leak configuration is required for these configurations. In this example, a VRF route leak configuration is required on the default VRF as follows:
  • ip route 100.1.1.2/32 next-hop-vrf blue <exit interface/next-hop-ip>
  • ip route 12.3.4.6/32 next-hop-vrf red <exit interface/next-hop-ip>
For inter-VRF deployment, use the use-vrf vrf-name option with the ip dhcp relay address command, where vrf-name is the VRF where the DHCP server is located.
Note

Note

As a best practice, do not configure the same DHCP relay address on different VRFs. For example:
9036691-01 Rev AA