The default configuration uses HTTP Basic authentication to challenge requests for the /xmla path. If client doesn’t have a valid JasperReports Server user name and password in its XML/A connection source, the connection will fail, unless the user name and password are left blank; in this case, the user name and password used to access the client application is passed to the server.
|
|
With HTTP Basic authentication, clear-text passwords are transmitted in the header of an HTTP request unless you have enabled JasperReports Server to use encrypted passwords. For more information, refer to the JasperReports Server Administrator Guide. |
When creating an XML/A connection, you can either specify a user name and password for all users to share, or you can leave user name and password blank, so that the connection passes the current user’s name and password to the server.
For a simple application with a relatively small dataset for which performance is not critical, (for example, a small intranet application or developer environment), the simplest server configuration is to run a single application server that connects to a local database, which contains both the data being analyzed (that is, the operational data store) and the JasperReports Server repository database. However, this configuration doesn’t perform under a large load. This section describes steps you can take to improve performance for larger implementations and data sets.
For larger analysis applications, the ROLAP database is often a performance bottleneck. In this case, dedicate a computer to host the ROLAP database, and configure JasperReports Server to access it. The database and its host should be optimized for disk read and write. This separation of the application from the data is sufficient for many implementations.
To improve performance further, separate the Jaspersoft OLAP user interface from other Jaspersoft OLAP elements. Inside JasperAnalysis, the application that visualizes your results and provides navigation (JPivot) competes for resources with the analysis engine (Mondrian). To separate these processing resources, run two different instances of JasperAnalysis: one that handles JPivot’s tasks, and another that handles Mondrian’s. The former instances is called the XML/A client; the latter is called the XML/A provider (or XML/A server). In the XML/A provider, define Mondrian connections that point to your operational data store. Then define XML/A connections in the XML/A client that point to Mondrian connections in the
XML/A provider.
In this configuration, your users connect to the XML/A client, which sends XML/A requests to the XML/A provider, which in turn connects to the database (using a Mondrian connection) to retrieve the results; it returns the results to the XML/A client. The XML/A client then provides visualization and navigation for your end users.
In this configuration, Mondrian and JPivot don’t compete for resources, and performance in large implementations is improved. For more information, see XML/A Configuration.
