Creating a Traffic IP Group

To create a traffic IP group, click Services > Traffic IP Groups. On this page, you can create new traffic IP groups and edit existing ones.

Enter a name and select the Virtual Traffic Managers you want to be members of the group. Enter the traffic IP addresses for the group in a list separated by spaces or commas, choose the IP distribution mode if necessary, and click Create Traffic IP Group.

Traffic Distribution

The IP distribution mode determines how Virtual Traffic Managers in the same Traffic IP Group share the traffic between them.

You can configure traffic IP groups to raise traffic IP addresses on one Virtual Traffic Manager at a time (single-hosted mode), or on all Virtual Traffic Managers in the group simultaneously (multi-hosted mode). On some product variants, you can also select Route Health Injection as the distribution mode.

Single-Hosted Mode

Each IP address in the group is only raised on one Virtual Traffic Manager at a time. If the group contains more than one IP address, they are dispersed across the participating Virtual Traffic Managers. When a Virtual Traffic Manager fails, its traffic IP addresses are moved to the other working Virtual Traffic Managers in the group.

Set the keeptogether option for a traffic IP group to prevent dispersion of the IP addresses in the group. In this case, all traffic IP addresses are raised on the same Virtual Traffic Manager. This is useful when using traffic IP groups with IP Transparency. For details, see Using IP Transparency with a Cluster.

Multi-Hosted Mode

Each IP address is raised on every machine at the same time. Incoming data for each IP is received by every Virtual Traffic Manager in the group, and each one takes responsibility for an equal portion of the incoming connections. If a machine fails, its portion is shared between the remaining active machines.

Multi-hosted IP addresses work by using multicast packets on your local network. The Virtual Traffic Manager advertises the traffic IP address using a multicast MAC address on the local subnet. This informs switches and routers that the traffic to the IP should be sent to all of the Virtual Traffic Managers. The MAC address used is calculated by providing a separate multicast group (represented as an IP between 239.0.0.0 and 239.255.255.255). This multicast group should be unique on your network.

Each Virtual Traffic Manager receives all incoming connections and makes a calculation based on the source IP address of the connection to determine whether it should accept that connection or if it should silently discard the connection (because another Virtual Traffic Manager in the group will accept it).

By discriminating on the source IP address alone, this method arranges that the same client will be processed by the same Virtual Traffic Manager (providing the client does not change IP address). This has the advantage that any session persistence data and SSL session data stored in the Virtual Traffic Manager will be available.

However, this can result in uneven distribution of traffic between Virtual Traffic Managers, particularly if your clients come from a small set of source IP addresses. This uneven distribution may impact the performance of the load-balanced service. If this is of concern, set the consider source port setting to instruct the Virtual Traffic Managers to distribute traffic based on the source IP address and port; this will share traffic much more easily across the cluster.

Note that using the source port may interfere with some session persistence methods, as there can be a short delay as the session persistence data is shared across the cluster (see Using Session Persistence with Multi-Hosted Traffic IP Addresses). Similarly, consider source port should not be used with any Traffic IP Addresses that are used by FTP virtual servers (see FTP).

If a Virtual Traffic Manager attempts to connect to a multi-hosted IP, the connection is always picked up by the local Virtual Traffic Manager itself.

Route Health Injection

The Virtual Traffic Manager uses front-end fault tolerance by announcing the Traffic IP addresses managed by the group out to the wider routing network. To use this method, the routers in your organization must be configured to use one of the following routing protocols:

•OSPFv2: Open Shortest Path First, version 2

•BGP: Border Gateway Protocol

You can only use RHI for IPv4 addresses; IPv6 is not supported.

Traffic IP Groups based on Route Health Injection (RHI) contain a maximum of two Virtual Traffic Managers, either with a single active Virtual Traffic Manager, or in a pair using an active-passive configuration.

When you select RHI as the distribution method, you additionally specify routing metrics for the active and passive Virtual Traffic Managers. When the Virtual Traffic Manager injects routes into the upstream routing network, these metrics determine the priority of the Virtual Traffic Manager instance within a multi-cluster (or multi-datacenter) scenario.

For more information, see Route Health Injection and the Network.

Passive Machines

If you mark a Virtual Traffic Manager as passive then it will not raise any of the IP addresses (single-hosted mode) or handle any load (multi-hosted mode) in the Traffic IP group unless one of the non-passive Virtual Traffic Managers has failed.

When you add a Virtual Traffic Manager to an existing Traffic IP Group, some of the traffic may be transferred to the new Virtual Traffic Manager. Unavoidably, this will result in some dropped connections at the instant of transfer. To deal with this situation, you can add a Virtual Traffic Manager as passive. In that case, it will not take any traffic unless one of its peers was to fail.

Disabling a Traffic IP Group

You can disable a Traffic IP Group by de-selecting the enabled checkbox. Use this option to disable a group temporarily, but to retain the configuration for future reinstatement.

When a group is disabled, none of the IP addresses in the group are raised. Route advertisements for RHI traffic IP groups are withdrawn.