Your Semaphore installation

Windows Services

Generally, the following Windows services exist on a Semaphore server:

• Classification server (“Semaphore <Version> - Classification Server <Revision>”) - This is the Classification Server application.
• Studio (“Semaphore <Version> - Studio (studio)”) - The master Semaphore Studio application that provides access to the “Knowledge Model Management”, “Document Analyzer” and “SES API” (Widgets) sub-components via the TCP/IP ports these separate sub-components listen on. This runs as a web application in a built-in Apache Tomcat instance (which runs within a Java VM). Note that even if this service is running, if the associated “da” and “kmm” services are not also running, these sub-components will not be accessible.
• Document Analyzer (“Semaphore <Version> - Studio (da)”) - The “document analyzer” sub-component of Semaphore Studio. This runs as a web application in a built-in Apache Tomcat instance (which runs within a Java VM).
• Knowledge Model Management (“Semaphore <Version> - Studio (kmm)”) - The “Knowledge Model Management” (model update) sub-component of Semaphore Studio. This runs as a web application in a built-in Apache Tomcat instance (which runs within a Java VM).
• Semantic Enhancement Server (“Semaphore <Version> - Semantic Enhancement Server”) - This is the Semantic Enhancement Server component of the Publisher-SES application.

These services can be started/stopped, as with all other Windows services, using the “Services” option in the Control Panel (“Administrative Tools”). Care should be taken to ensure that all of these services are set to re-start on server reboot (“Startup type” of “Automatic”).

Scheduled tasks

Scheduled tasks are not created by a standard Semaphore installation. Any scheduled tasks created will have been created for site-specific requirements such as model backup or for scheduling the publishing process.

Firewall considerations

There are a small number of TCP/IP ports used by Semaphore. By default, for the Semaphore server the following ports must be open externally:

• 5080 - Port for users to be able to access Semaphore Studio

The sub-components of Studio also listen on TCP/IP ports though they are generally not accessed across the network (and are installed as part of Semaphore Studio):

• 5081 - Default port for “Document Analyzer”, Semaphore Studio sub-component
• 5082 - Default port for “Knowledge Model Management”, Semaphore Studio sub-component
• 5083 - Default port for “Rulenet Manager”, Semaphore Studio sub-component
• 5084 - Default port for “SES Manager”, Semaphore Studio sub-component

There are additional ports used for components to talk to one another. If the applications are partitioned, that is, installed on different machines then these ports will also need to be open:

• 5058 - Classification Server integrations wishing to communicate with the Classification Server API (programmatically).
• 5059 - Classification Server administrators (users of the “Classification Analysis Tool”) must be able to communicate via this port (this is set in the Classification Server configuration file by default this is “config.xml”). It is not generally required that this port be open to external clients unless they wish to use the test interface.
• 8983 - Semantic Enhancement Server clients (for example those applications wishing to access Semantic Enhancement Server functionality).
• 9983 - Publisher will need access to this Semantic Enhancement Server port (if it is running in the default “SOLR Cloud” mode)

See Appendix - Detailed Network Firewall Rules for further information.

Windows desktop security considerations

The Text Miner client software must be installed by a system administrator or other privileged user.

If the user is to communicate with Classification Server test interface, the user’s local network settings (e.g. Internet connection settings/proxy server) should support connecting to the relevant computer (server) via the configured port (e.g. 5058 or 5059).