You use a number of machines in this pattern. The purpose of each machine is described in detail below.
Do not follow the pattern in the order shown below, as this overview of the system landscape describes the role of each machine in this pattern; it does not describe the order to follow. To follow the pattern, complete the tasks in the order shown in Pattern Overview.
The following diagram shows the machines in this pattern and how they connect with one another:
External machine (
The external machine hosting the web server, reverse proxy server, and load balancer is typically outside the company's external firewall. Customers can access BI platform web applications, such as BI launch pad and the Central Management Console (CMC), from the client machine.
We chose Apache because it is the most commonly used web server on the internet, hosting over 60% of all websites.
Apache 2.2 is the current mainstream release and is the latest supported version of Reverse Proxy Server in BI 4.0 Feature Pack 3.
See Apache for setup details.
Application servers (
The application servers host BI platform web applications, such as BI launch pad and the CMC. For production systems, you should deploy at least two web application servers to handle failover and load balancing.
For this pattern, we chose Tomcat 7.0 for the application server rather than Tomcat 6.0, which is included in the BI platform installation program. Although Tomcat 6.0 is acceptable for production deployments on Linux, we want to show how to deploy BI platform web applications to the company's own application server.
See Application Server for setup details.
Authentication server (
All users are authenticated before they can access BI platform. Typically, a company integrates BI platform with its own authentication system such as LDAP or Windows AD. Although BI platform includes its own authentication system by Enterprise authentication, it is not typically used in production deployments because it is less secure than third-party authentication systems.
For this pattern, we chose AD authentication although other third-party systems could also be used.
See Authentication Server for setup details.
Clustered BI platform servers (
BI platform servers handle requests from the application servers, retrieve data from the corporate and system databases, generate BI content, and return the content to the application servers.
This pattern uses a two-machine cluster, in which each machine (
lnxpb05) hosts all of the BI platform servers. This configuration provides redundancy and some scalability. BI platform web applications detect both BI platform server machines (specifically, they detect all CMS servers in a cluster), and automatically load balance their requests. If one machine fails, the web applications automatically send all requests to the other machine.
For production systems, you should deploy at least two BI platform server machines. If you need to increase scalability, you can deploy more BI platform server machines and deploy only one or two BI platform servers per machine.
See BusinessObjects Cluster for setup details.
File repository server (
The file repository server (FRS) contains all of the reports and other BI documents that have been created. When a user creates a BI document, the BI platform servers retrieve the data from the company database (Reporting DB), generate the document, and store it in the file repository server. In production environments, you should host the FRS on its own machine because the number of BI documents can grow to a large number.
Because BI documents often contain sensitive information, you should host the FRS on a secure machine. Typically the BI platform administrator does not have access to this machine.
BI platform system database
BI platform uses a database for its system data, which is called the system database or CMS database. For example, when a user wants to see a report, the CMS checks the system database to find the location of the report. For production systems, you can host the CMS database either on the database included with BI platform or on the company's existing database. Both are recommended for production systems, but companies often use their own databases in order to simplify administration.
We chose to use Sybase ASE because it is a commonly used database in Linux deployments.
See CMS Database (Linux) for setup details.
The reporting database hosts the company’s data. When a user wants to report from, analyze, or view company data, the BI platform servers retrieve data from the company’s database, perform the analysis and formatting, and either serve the result directly to the application layer or store it in the FRS.
See Reporting Database (Linux) for setup details.