6.2.1 Using the in-process HTTP server
The in-process HTTP server is the Web server functionality provided in the J2EE server processes.
As the J2EE server processing receives the HTTP request directly without passing through the Web server, you can use the Web server functionality with even better processing performance than during the Web server integration. Therefore, for a system that emphasizes performance, Hitachi recommends that you use the in-process HTTP server.
However, there are comparative differences in the functionality provided in Cosminexus HTTP Server and Microsoft IIS, you check the differences in the functionality, and then determine whether to use the in-process HTTP server. Furthermore, you cannot use the in-process HTTP server and the Web server integration functionality simultaneously. When you design the system, you must choose which functionality you will use, in advance.
To use the in-process HTTP server, the prerequisites are as follows:
-
The in-process HTTP server must be deployed within the internal network in failover instead of deploying the server in DMZ that is accessible from the external networks to which unauthorized access is assumed. In a system accessed from external networks such as Internet, you must build a system in which a proxy server is deployed on DMZ and forwarded to the in-process HTTP server within the internal network.
-
In the in-process HTTP server, only HTTP is supported. HTTPS is not supported. To use HTTPS, the SSL accelerator or reverse proxy of Cosminexus HTTP Server is a prerequisite.
-
You can use the in-process HTTP server to access only the Web applications deployed on the J2EE server. Note that you cannot deploy static contents alone, but only when you execute request distribution with the redirector and error page customization, you can specify static contents that are not included in the Web application.
Take note of the following when using the in-process HTTP server:
-
If you stop the J2EE server by executing the cjstopsv command, during the TCP connection of the Web client and the in-process HTTP server, the J2EE server does not stop, until the Web client disconnects the TCP connection of the in-process HTTP server or the timeout specified in the webserver.connector.inprocess_http.persistent_connection.timeout key of usrconf.properties (user property file for the J2EE server) occurs. If you want to stop the J2EE server regardless of the disconnection of the TCP connection from the Web client or the timeout occurrence, forcibly stop the J2EE server by specifying the -f option in the cjstopsv command.
Customize the J2EE server properties to specify the settings for the in-process HTTP server. For details on customizing the operation settings for the J2EE server, see 6.2.3 Execution environment settings (J2EE server settings).
- Tip
-
The in-process HTTP server is not the default server.