Understanding How ARP Works

Figure 1 and Figure 2 show how ARP works where host 1 sends an IP packet to host 2 on a different subnet. To complete this transmission, host 1 needs the MAC address of router 1, to be used as the forwarding gateway.

A typical scenario is:

1. Host 1 broadcasts an ARP request to all devices on subnet 1, composed by a query with the IP address of router 1. The IP address of router 1 is needed because host 2 is on a different subnet.
2. All devices on subnet 1 compare their IP address with the enclosed IP address sent by host 1.
3. Having the matching IP address, router 1 sends an ARP response, which includes its MAC address, to host 1.

   Figure 1: Sample ARP Process—1 through 3

   
4. Host 1 transmits the IP packet to layer 3 DA (host 2) using router 1’s MAC address.
5. Router 1 forwards IP packet to host 2. Router 1 might send an ARP request to identify the MAC of host 2. (See Figure 2.)

   Figure 2: Sample ARP Process—4 and 5

   

ARP forces all receiving hosts to compare their IP addresses with the IP address of the ARP request. So if host 1 sends another IP packet to host 2, host 1 searches its ARP table for the router 1 MAC address.

If the default router/gateway becomes unavailable, then all the routing/packet forwarding to remote destinations ceases. Usually, manual intervention is required to restore connectivity, even though alternative paths may be available. Alternatively, Virtual Router Redundancy Protocol (VRRP) may be used to prevent loss of connectivity. See JunosE IP Services Configuration Guide.