ARP Load-Sharing

In a Layer 3 interface deployment and active/active HA configuration, ARP load-sharing allows the firewalls to share an IP address and provide gateway services. Use ARP load-sharing only when no Layer 3 device exists between the firewall and end hosts, that is, when end hosts use the firewall as their default gateway.
HA_ARP_load_sharing.png
In such a scenario, all hosts are configured with a single gateway IP address. One of the firewalls responds to ARP requests for the gateway IP address with its virtual MAC address. Each firewall has a unique virtual MAC address generated for the shared IP address. The load-sharing algorithm that controls which firewall will respond to the ARP request is configurable; it is determined by computing the hash or modulo of the source IP address of the ARP request.
After the end host receives the ARP response from the gateway, it caches the MAC address and all traffic from the host is routed via the firewall that responded with the virtual MAC address for the lifetime of the ARP cache. The lifetime of the ARP cache depends on the end host operating system.
If a link or firewall fails, the floating IP address and virtual MAC address move over to the functional firewall. The functional firewall sends gratuitous ARPs to update the MAC table of the connected switches to redirect traffic from the failed firewall to itself. See Use Case: Configure Active/Active HA with ARP Load-Sharing.
You can configure interfaces on the WAN side of the HA firewalls with floating IP addresses, and configure interfaces on the LAN side of the HA firewalls with a shared IP address for ARP load-sharing. For example, the figure below illustrates floating IP addresses for the upstream WAN edge routers and an ARP load-sharing address for the hosts on the LAN segment.
HA_mixed_float_and_ARP.png

Related Documentation