5.9.6 Distributing Materialized Views (Data Grid)
Report Server, Data Server, and Data Node on Same Machine (Default)
Report and Data Server on Machine 1, Data Nodes on Machines 2-N
Report Server on Machine 1, Data Server on Machine 2, Data Nodes on Machines 3-N
Default Materialization in a Clustered Environment
To improve Viewsheet performance, you can distribute a materialized view across multiple machines. This distributed configuration is called a data grid. In a data grid, the materialized dataset is saved on separate physical machines (nodes) in the grid. This allows the Viewsheet server to pull small blocks of data from each of the nodes, thereby distributing the processing load and minimizing hard disk access for any particular machine. This increases total throughput and improves Viewsheet responsiveness.
A data grid is completely transparent to the end-user. Users can access their Viewsheets in the same manner regardless of whether a data grid has been configured or not. However, Viewsheet performance will in general be better with a data grid in place.
A data grid configuration consists of the following components:
• Report Server: The Report Server is a J2EE web application that handles all user requests for reports and Viewsheets in the repository. The Report Server also provides access to the Enterprise Manager for managing the repository and the Scheduler application. For Viewsheets that utilize materialized views, the Report Server delegates data retrieval to the Data Server.
• Data Server: The Data Server is a J2EE web application that handles materialized view requests from the Report Server and dispatches tasks to the Data Nodes. The Data Server also provides access to the Enterprise Manager for monitoring data grid status and managing materialized views. The Data Server requires an 'H' or 'G' key.
Note: Each Data Node must be a distinct machine.
• Data Node: A Data Node is a machine running the Data Node service, a stand-alone Java process that has no user interface and does not require a web application container. The Data Nodes store data from the materialized view and respond to requests for data from the Data Server, (via socket connections). A Data Node requires an 'H' or 'G' key.
You can configure the Report Server, Data Server, and Data Node in various ways, as described in the following sections. Note that the Data Server and a Data Node cannot reside on the same machine, except in the default installation (Report Server, Data Server, and Data Node on Same Machine (Default)).
| << 5.9.5 Managing Materialized Views | © 1996-2013 InetSoft Technology Corporation (v11.4) | Report Server, Data Server, and Data Node on Same Machine (Default) >> |