Distributed configuration

To promote scalability, server-side components can be distributed across several servers.

A distributed server configuration enables seamless load-balancing between multiple computers to accelerate response time and increase throughput. Distribution of the processes is flexible because the system does not require that any two components run on the same computer.

Start of change You can configure whether load balancing is performed by the listener or by read/write instances. End of change

Figure 1. Distributed configuration

A client and the Management Console connect to a server that hosts a listener, an administration process, a read/write knowledge base, a read-only knowledge base, a read-only decision plan, and the data server. The listener points to read/write and read-only knowledge bases and decision plans on another server. All instances point to the data server.

To set up this configuration:

Install the Content Classification server on computer A and select the option to run the administration and data server on this computer. Also, select the option to run the listener component on this computer.
Install the Content Classification server on computer B. Select the option to connect to the administration and data server on a remote computer and specify the host name and port of computer A. Clear the option to run the listener component on this computer.
In the Management Console, add the decision plan and referenced knowledge bases to the server and start them. When you add each decision plan and knowledge base, specify the computer on which to run the read/write process by configuring the Read/write server setting. To set up read-only processes on the two computers, configure the Read-only servers setting.
Specify whether load balancing is to be performed by the listener or by read/write instances. Configure the Routing Requests options in the Management Console Global Properties window. If you select the option to route incoming requests to the listener, ensure that the following option is selected for best performance: Route requests only to decision plans or knowledge bases on the server that received the request.