The main entry point to Nucleus is though its request-handling facilities. When a Web application running Nucleus receives an HTTP request, it invokes a series of servlets and servlet filters that process the request and construct the response:
When a user requests a page, that request is sent to the Web server as a stream of information that the Web server parses and holds in an
HTTPServletRequest
object.The Web server passes the
HTTPServletRequest
to the application server, which it wraps in its own flavor of request around the generic one before passing it to the Web application.If you defined custom filters and the request is for a JHTML page, the filters execute in the order you specified.
Any custom J2EE servlets are processed.
The Web application calls one of the following resources. Each of these resources starts a different request-handling pipeline:
DynamoProxyServlet
, called when the request is for a JHTML page, starts the DAS servlet pipeline.PageFilter
, called when the request is for a JSP, starts the DAF servlet pipeline.
The request is passed on to a pipeline of servlets. Each servlet available to the pipeline is designed to insert itself into the pipeline when the information it processes is included in the request. So, a unique pipeline is constructed for each request. If you created any custom Dynamo servlets and they apply to the current request, they are executed now.
If you created custom filters and a JSP is being requested, they execute after the last pipeline servlet, but before the request returns to the application server.
Once the servlet pipeline reaches the end, it returns the request to the application server for final processing (executing the servlet representation of the page).
The following figure illustrates this process, and shows how the processing differs depending on whether the request is JSP or JHTML:
![](media/image4.png)
For more information about ATG’s request handling pipelines, see the Request Handling with Servlet Pipelines chapter in this guide.