Typically, when your DHCP server is located under an EPG, all the endpoints in that EPG have access to it and can obtain the
IP addresses via DHCP. However, in many deployment scenarios, the DHCP server may not exist in the same EPG, BD, or VRF as
all the clients that require it. In these cases a DHCP relay can be configured to allow endpoints in one EPG to obtain IP
addresses via DHCP from a server that is located in another EPG/BD deployed in a different site or even connected externally
to the fabric and reachable via an L3Out connection.
You can create the DHCP
Relay policy in the Orchestrator GUI to configure the relay. Additionally, you can choose to create a DHCP
Option policy to configure additional options you can use with the relay policy to provide specific configuration details. For all
available DHCP options refer to RFC 2132.
When creating a DHCP relay policy, you specify an EPG (for example,
epg1) or external EPG (for example,
ext-epg1) where the DHCP server resides. After you create the DHCP policy, you associate it with a bridge domain, which in turn is
associated with another EPG (for example,
epg2) allowing the endpoints in that EPG to reach the DHCP server. Finally, you create a contract between the relay EPG (
ext-epg1) and application EPG (
epg2) to allow communication. The DHCP policies you create are pushed to the APIC when the bridge domain to which the policy is
associated is deployed to a site.