Z26688-02
 Contents |
| Oracle Discoverer 3i Viewer Edition Installation and Administration Guide Z26688-02 |
|
Oracle Discoverer 3i Viewer Edition is a business intelligence tool for viewing workbooks created by users of the Windows or Web releases of the User Edition. Using Oracle Discoverer's easy to use interface via a web browser, users can access and analyze database data.
Oracle Discoverer 3i Viewer Edition can be rapidly deployed by using existing Discoverer installations and is straightforward to set up and maintain. Oracle Discoverer 3i Viewer Edition can be used to integrate database output into your web site and portal and is easily customized to fit in with your web site look and feel, or to build custom Discoverer applications for the web. Oracle Discoverer 3i Viewer Edition scales up easily to support more users as demand on the system increases. It is also optimized for performance and designed to minimize network traffic.
This chapter provides an overview of the Oracle Discoverer 3i Viewer Edition architecture, describes the components and illustrates how Oracle Discoverer 3i Viewer Edition works to give data access to users.
Oracle Discoverer 3i Viewer can be used to view workbooks created by the Discoverer User Edition on windows or the Discoverer User Edition on the web. The deployment of Oracle Discoverer 3i Viewer Edition is best understood by looking at its context (See Figure 1-1).
.
Oracle Discoverer 3i Viewer Edition is deployed using an internet computing architecture that sits on top of the existing Oracle Discoverer Application Server.
Oracle Discoverer Application Server works with an HTTP server to provide a robust, scaleable deployment platform for the Oracle Discoverer 3i Viewer Edition.
The n-tiers of Discoverer are as follows:
This architecture takes advantage of the distributed nature of the Web environment. While it is possible to install all tiers of Discoverer on the same machine, we recommend distributing your installation over multiple machines to maximize performance.
The following figure illustrates the tiered architecture.
The first tier is the client--a user's computer. There is no setup or installation to do on the client machine, nor does this machine need to have access to any database. The only requirement for this computer is that it can run html through a Web browser, such as Microsoft Internet Explorer 4.01 or Netscape Navigator 4.0 (See Section 2.1.3.1, "Client Web Browsers"). By accessing the URL that you (the Discoverer administrator) provide, users can connect and run Oracle Discoverer 3i Viewer Edition to view data.
The second tier is the HTTP Server having the Discoverer 3i Viewer Servlet installed in the web server's servlet engine. The Locator must be accessible from the HTTP server.
A Servlet comprises modules of Java code that run as a server application to answer HTTP requests (hence the name "Servlets", similar to "Applets" which are used on the client side). The servlet reduces to a minimum any client-side processing.
A Servlet Engine comes as part of (or a plug in to) an HTTP Server and is the environment in which the servlet runs. The Servlet Engine incorporates a Java Virtual Machine (JVM) and implements the Java servlet API.
The HTTP server (servlet engine) contains the following components:
The Discoverer Servlet generates HTML pages from HTTP requests using an XSL processor.
The third tier is the Discoverer Application Server. The Server consists of three basic components: the Locator, the Session, and the Preferences.
The components are installed on one or more application servers. You can install Discoverer Application Server on one machine or across many machines.
In addition, two CORBA (Common Object Request Broker Architecture) components from VisiBroker--an Object Activation Daemon (OAD) and a SmartAgent component--are also installed. These CORBA components are responsible for activating new components and providing a location service when a user requests that a Discoverer session begin.
The Discoverer Application Server contains the following components:
A Discoverer server session is created for each user and performs all the business intelligence of Discoverer 3i Viewer for a given user. Once a session is started, it is the user's instance of Discoverer 3i Viewer. The Session component contains all of the application logic and performs all of the Discoverer operations such as connecting to the database or opening a workbook.
Machines with a Session component installed can run multiple concurrent sessions for clients. To increase overall performance, you can install the Session component on many individual application servers, and each one of them can run multiple concurrent sessions. Distributing the Session component on several servers also improves availability. If one server is not operating, other servers handle the session requests.
The Session component can run on any server (application or HTTP) in the Discoverer Server configuration as long as the server is a Windows NT machine.
The purpose of the Locator is to receive a request for a session from the servlet (the request is initiated after a URL is sent to the servlet from the client browser), find the next application server available, start a new session, and return a reference for the new session back to the servlet. Once the servlet has received this reference, the servlet and the session communicate via the Discoverer Server Interface(the model); and the Locator waits for the next request. The Locator component is installed on the HTTP server. You need only one Locator component on the network for the Discoverer Server.
The Locator component also helps balance the load on the available application servers because it determines which application server will start each requested session. You can specify how the Locator allocates sessions to achieve the best performance.
The Preferences component provides a single location for preference settings for all end users. The Discoverer Server relies on stored preference settings to dictate certain aspects of behavior. You install only one Preferences component for the Oracle Discoverer Server.
Having a single Preferences component is important in a distributed environment where different components can run on different machines. The Preferences component provides consistent preference settings for all components regardless of where they are running. The Preferences component stores and reads preferences from the Windows Registry. The Preferences component must be installed on a server running Windows NT.
The fourth tier of Discoverer is the database. The database contains the data users want, the End User Layer (EUL) that provides an easy to understand view of the data, and the Discoverer workbooks that users can work with to view and analyze the data. The EUL is created and maintained with the Discoverer Administrator Edition.
The Discoverer components must each be installed on an appropriate machine. Three kinds of server can make up the Discoverer architecture:
The machine on which the HTTP server and servlet engine are installed (also hosts the Discoverer 3i Viewer Servlet component). If you install Discoverer Application server on a single machine, the machine must be an NT server (See Section 2.1.3.2, "HTTP Servers supported by Oracle Discoverer 3i Viewer Edition").
If you distribute the software components of the Discoverer Application Server across several machines, then install the Discoverer Application servers and the Session components on the same machine. You can have many servers as part of your Discoverer environment. For example, if you install the Session component on three machines, each is a Discoverer Application server and multiple sessions can be started on each one.
If you install all components on the same machine it becomes both the Discoverer Application server and HTTP server.
If your Discoverer Application Server environment includes several servers, you should designate one as the Master Discoverer Server. It should contain the Preferences and Session components. The remaining servers will only host the Session component.
This section describes how Oracle Discoverer 3i Viewer Edition works--from the point when the user launches the browser on the client computer to when data access becomes available. The following figure illustrates the basic process.
Oracle Discoverer 3i Viewer Edition is invoked via a URL from a standard Web Browser.
The URL is processed by a servlet running in the servlet engine. The servlet uses the Locator to communicate with the Discoverer Application Server.
The Oracle Discoverer 3i Viewer Edition Servlet interprets the HTTP request from the client browser, and makes the necessary calls to the Discoverer Application Server.
The server response represented in XML is generated by the servlet and sent to an XML/XSL processor.
This combines the XML with an XSL configuration file that defines the representation of the User Interface and generates the output(eg HTML) to send back to the browser.
The Oracle Discoverer 3i Viewer Edition is composed of two primary layers - the application logic and the presentation logic.
The application logic manages the handling of the HTTP request, the state of the Discoverer application server, and generates the XML representing the logical state of the Discoverer server and the data for a given client.
The presentation logic takes the XML and uses XSL to generate a response in a given format (e.g. HTML, Excel, XML etc.). The presentation logic can be customized to alter the appearance of user interface(by editing XSL files). See Section 5, "Customizing Oracle Discoverer 3i Viewer Edition" for details.
The Oracle Discoverer 3i Viewer Edition Servlet and Discoverer Application Server can both run on the Web Server, if required (See Section 2.3.1.4, "Installing Oracle Discoverer 3i Viewer Edition Component and Master Server on a Single Machine"), but they are designed to be deployed on separate and multiple machines (See the Oracle Discoverer 3i Installation and Administration ). The Web Server may then be replicated using standard Web Farm techniques and the Discoverer Application Server can be replicated by specifying the machines available to Discoverer in a configuration file. Discoverer Application Server Sessions are then started on alternate machines to balance the load.
In this way very large numbers of Discoverer Users may be supported, since there is no single bottleneck. Multiple Discoverer sessions can run on a single machine, or can be spread across multiple machines as required.
The Discoverer 3i Viewer servlet can also be distributed.
In Figure 1-3 two web servers are used to spread the load for three browser sessions. There are three server machines, with one Discoverer Application Server running on each. In a real system there would be many users using each web and application server. Discoverer allows you to determine exactly how you want to spread the load across the available machines.
|
 Copyright © 2000 Oracle Corporation. All Rights Reserved. |
|