4.7.2 Load balancer
You can use the load balancer to improve the processing performance by distributing traffic efficiently. The load balancing functionality controls multiple servers with a single virtual IP address. You can use BIG-IP (BIG-IP v9, BIG-IP v10.1, BIG-IP v10.2, or BIG-IP v11) or ACOS (AX2000, AX2500, BS320) as load balancers.
You can use the Cookie switching functionality of the load balancer to operate the system set up with the Smart Composer functionality. The cookie switching functionality maintains sessions by using the cookies. Even if you cannot identify the client from the source IP address via proxy, you can caste a series of HTTP requests from the same client to the same service unit.
- Reference note
-
Switching mode
-
In 08-53 or later versions, only the Insert mode among the Cookie switching function of BIG-IP is supported as a switching mode by the Smart Composer functionality.
-
ACOS sets up cookies against the responses from the server. Format of cookies to be set is as follows:
cookie-name-virtual-port=real-server-IP-address_real-port
-
Also, in a system set up by using the Smart Composer functionality, you can share the load balancer among different Web systems or among the same Web system having different management domains, or allocate the load balancer with redundant configuration to a single Web system.
- Important note
-
You cannot deploy the load balancer if you use the free-tier pattern or the pattern that does not use Web server integration (j2ee-tier).
- Organization of this subsection
(1) Controlling the load balancer
Use the CLI command or APIs to control the load balancer from Application Server. For that, set the connection information of the load balancer in the lb.properties (load balancer definition property) file. For details on the procedure to set the connection information, see 4.7.5 Setting up environment for connecting to the load balancer.
(2) Sharing the load balancer among Web systems
You can share same (single) load balancer among different Web systems set up by using the Smart Composer functionality. Note that when sharing a single load balancer among different Web systems allocated to a single host, you must separate the virtual server of the load balancer for each Web system.
For notes on sharing a load balancer, see 4.7.6 Notes on controlling the load balancer.
(3) Load balancer redundancy
In the Web systems set up using Smart Composer functionality, you can allocate the load balancer with redundant configuration (active and standby configuration) in a single Web system. You can allocate maximum two load balancers. If you allocate the load balancer with redundant configuration, and the running system (active) of the load balancer fails, request casting automatically switches to the standby system (standby), and the operation continues. For notes on the load balancer redundancy, see 4.7.6 Notes on controlling the load balancer.
- Reference note
-
When you want to change the load balancer of the existing Web system to the load balancer with redundant configuration, you must first delete the existing Web system, and then re-set up the Web system with the load balancer redundancy.