Load balancer

A load balancer distributes workloads across two or more servers, network links, or other resources based on criteria defined in the load balancer configuration.

Virtual server

A virtual server is an IP address (virtual IP) and service port combination, for example 172.16.200.210:80. A virtual server’s main function is to distribute traffic loads to a group of internal, back-end servers. All traffic is directed to the virtual server, and the server distributes the traffic across the back-end devices.

Virtual servers are configured differently on Brocade ADX load balancer and F5 BIG-IP load balancer devices.

Brocade ADX load balancer virtual server configuration

On an Brocade ADX load balancer, virtual servers are defined explicitly by binding virtual servers with real server resources. A real server is a configuration object existing within the ADX that defines the IP address of the internal, back-end server. You associate server resources with a virtual server by using the virtual server binding syntax as shown in the following example:

Logical component	Configuration
Virtual server	server virtual VS-1.2.3.4 172.16.200.210
real server1 binding	bind http web1 bind http web2
real server2 binding	bind http web2

Note

If a real server has multiple service ports configured, you can use those ports as resources for the virtual server.

F5 BIG-IP load balancer virtual server configuration

On a F5 BIG-IP load balancer, you configure the backend devices associated with a virtual server by using pools that specify the configuration for the backend devices and pool members for managing traffic.

Logical component	Configuration
VIP	VS-1.2.3.4:80
Pool	POOL-172.16.200.210-80
Member 1	10.200.10.25:80
Member 2	10.200.10.26:80
Member 3	10.200.10.27:80

BIG-IP F5 load balancer configuration

After a pool is applied to a virtual server, client traffic destined for the virtual server uses the pool as its resource. For a pool to serve its function of fault tolerance and redundancy, multiple pool members must be available inside the pool. When a pool has multiple members, the load balancer decides which one receives the client traffic based on the load balancing method configured on the device (such as Round Robin, Weighted Robin, or Least Connection).

Pools

On F5 BIG-IP load balancer, a load balancing pool is a configuration object created in the Local Traffic Management system. Each pool is a logical grouping of devices used to receive and process traffic.

You can define multiple pools in the F5 BIG-IP load balancer configuration. However, you associate each pool with only one virtual server that distributes traffic across the devices in the pool.

After you create a pool, you configure the backend resources in the pool by adding pool members.

Pool member

A pool member is a logical object representing a single internal physical server IP address and listener port. You can include a pool member in multiple pools.

Node

A node is a logical object that provides the IP address of a single, physical backend device, such as a web server. Nodes are the base configuration object when creating a virtual server.

In an F5 BIG-IP load balancer, the system automatically adds the node when you add a member to a pool.

Event

You can use events to track asynchronous backend processing status. The system generates these events when a user performs an action on the load balancer that updates a device. The returned response object includes an event ID, event type, status of the request, and the timestamp when the system created the event.

Monitor

For a F5 BIG-IP load balancer, monitors verify the health and availability of a node, pool, or a group of nodes in a pool.

A health monitor, or a health check, is a configuration object that specifies a check and parameter values to verify the health and status of a load balancer component such as a pool. Checks run continuously at a time interval specified in the monitor configuration. If the component does not reply correctly to the health monitor before the timeout value expires, the system identifies it as having degraded performance and removes that component from the availability pool.

A monitor does not take effect until you apply it to a virtual server, a pool, or a pool member. You apply it by submitting an API request to create a monitor rule. After applying a rule, you can use the update monitor rule operations to change configuration settings.

On the F5 BIG-IP load balancer, the default interval timer is five seconds, with a default timeout value of 16 seconds.