Help us improve your experience.

Let us know what you think.

Do you have time for a two-minute survey?

TechLibrary
Navigation

MX Series Router Cloud CPE VRRP and DHCP Services Overview

For customers running VRRP between two cCPE instances for dual access and who also want to use DHCP to allocate addresses for their devices, you must determine the best place to run the DHCP server.

The first option is running two DHCP servers, one in each cCPE instance, and splitting the address pool between the two servers. In this scenario, both DHCP servers are active and devices can receive an IP address from either server. Both DHCP servers configure the VRRP virtual address as the default gateway, so that subscriber traffic travels through the active VRRP interface. When the active access link goes down, the DHCP server in the corresponding MX Series router is unreachable to user devices. New DHCP leases are provided by the live DHCP server; however, DHCP renew requests for the leases managed by the unavailable DHCP server will fail. Devices requesting renewals receive new addresses from the live server. Existing TCP connections to or from the device are broken, and the application might need to be restarted to recover.

The second option, shown in Figure 1, is to run stateless DHCP relay agents in each cCPE instance. Both DHCP relay agents must point to the same DHCP server. When a device sends a DHCP discovery request, the request is forwarded to the DHCP server by both relay agents. The DHCP server may send two offers back to the client through the relay agents; however, the client will select an offer and send a request for it. The DHCP server sends an acknowledgement back to the client and the other offer expires. DHCP renew requests are sent to the sever using unicast. When one of the cCPE instances is unavailable due to the corresponding access link being down, DHCP discover requests are sent to the DHCP server by the other relay agent without interruption. Existing DHCP leases are renewed successfully because renew requests are sent by means of unicast to the DHCP server without relay agent involvement.

Note: The stateless DHCP relay agent, which you can configure under the [routing-instance routing-instance forwarding-options helpers bootp] edit hierarchy, cannot coexist with the DHCP server in the same MX Series router, even if they are configured in different routing instances.

If one or more DHCP server instances are running on the PE router, you must configure the DHCP relay agent by using the [routing-instance routing-instance-name forwarding-options dhcp-relay] configuration statement. However, when you use this relay agent, it is possible that renew requests can be dropped if the router has not detected the original discovery request for the address. If the renew request is dropped by the relay agent, the client is forced to send a DHCP discover request when its lease expires. Most likely, the client will receive the same IP address again.

Figure 1: MX Series Router Cloud CPE VRRP with DHCP Relay Service

MX Series Router Cloud CPE VRRP with DHCP Relay Service
 

Related Documentation

 

Modified: 2015-11-05

Previous Page Next Page