Applies To:

Show Versions Show Versions

Manual Chapter: Introduction to All-Active Clustering Using ECMP
Manual Chapter
Table of Contents   |   Next Chapter >>

Overview: Configuring all-active clustering using ECMP

You can implement BIG-IP® high availability on a network by using the Equal Cost Multi-Path (ECMP) protocol. You do this by creating the same virtual IP address on each BIG-IP device in a Sync-Failover device group. Then the upstream router uses ECMP to determine if there are multiple, equal-cost paths to this virtual address. If there are, the router uses an algorithm to select a path to the external floating self IP address on any one of the BIG-IP devices.

With this configuration, an application is not pinned to a specific device; instead, all of the devices actively process the same application traffic. And you can add capacity at any time by adding a BIG-IP device to the device group, syncing the configuration to the new device, and then modifying the configuration on the upstream router.

This illustration shows an example of this configuration. Notice that each device in the device group has the same virtual address.

BIG-IP system clustering using ECMP

High-availability clustering using ECMP

The illustration shows that the floating virtual address has been configured to span all three BIG-IP devices. Each device also has an active traffic group containing a unique floating external self IP address and a unique floating internal self IP address. So the floating addresses for BIGIP_A are assigned to traffic-group-1, the addresses for BIGIP_B are assigned to traffic-group-2, and the BIGIP_C addresses are assigned to traffic-group-3.

On the upstream ECMP-enabled router, routes are configured to send packets for App_1 to any of the three floating external self IP addresses. If an active traffic group fails over to another device, the packets targeting that traffic group's floating self IP address begin to target the same traffic group on the newly-active device. So for example, if traffic-group-1 fails over to BIGIP_B, then any traffic destined for the external floating IP address now targets the same address on BIGIP_B.

The floating internal self IP address within each traffic group ensures that on the return trip from a server, the packets go back through one of the BIG-IP devices. Using the SNAT Auto Map feature, a return packet will always target one of the floating internal self IP addresses in the device group.

Before you begin

Use this information to make sure that you have configured the prerequisites for this ECMP implementation.

Static route configuration
On the upstream ECMP-enabled router, make sure you configure a route to each floating external self IP address in the device group (one per device).
DNS servers (optional)
If you intend to use fully-qualified domain names (FQDNs) for either your server pool members or your NTP servers, you need to specify one or more DNS servers on each device.
Table of Contents   |   Next Chapter >>

Was this resource helpful in solving your issue?

NOTE: Please do not provide personal information.

Additional Comments (optional)