Features

ETL-Architektur

ETL:

ETL stands for extract, transform and load and describes the loading process of heterogeneous data from several source systems into a target system.

Drill-Through:

Drill-Through describes the linking from one data source to another, most often a more detailed data source. Palo ETL Server facilitates the Drill-Through from Palo to the document data in the source systems and provides information on the data’s origin at a very high detail level. The required data of the source systems is cached in Data-Sources of ETL-Server.

Connections:

All common relational database systems such as MySQL, DB2 and Oracle for example, can be connected via a JDBC driver. Of course, the read and write of flat files is supported as well. A high-performance interface exists for Palo Server as both source and target system. Further, a direct interface to SAP systems via RFC/BAPI is being planned for a future release.

Data-Sources:

Data is loaded from the relational source systems into Data-Sources via any SQL statement. When reading data from Palo Server, regular expressions can be used in the filter. Furthermore, Data-Sources allows a join of data from various source systems. Generic models of dimensions, e.g. periods and value types, can be created directly on the ETL-Server without referencing a source system.

Pipelines / Transformers:

Pipelines transfer a Data-Source into another Data-Source. At the same time, data may be cleansed, filtered and aggregated. Levels, attributes and weights for dimension data can be created directly. Thus, a relational model turns into an OLAP model in an easy way.

Transformers carry out special conversions on field level. Pre-defined transformers, for the purpose of mapping, data-chaining and date for example, are available. Customized transformers can be implemented in Java to cover more complex requirements.

Exports / Jobs:

The Data-Sources are loaded into the target systems via the help of Exports. Therewith, a complete Palo model, including dimensions and cubes, can be created directly on ETL-Server.

Jobs group several exports. They are the basis for a time-controlled and fully automated load process. The Client-Server architecture warrants a high scalability.

User Interface:

The ETL process is defined in a clear XML structure. Eclipse IDE for example is a suitable User Interface for the development process. The integration of the load process in external scheduling tools is possible. A Graphical User Interface is scheduled to be released in 2008 along with multi-user functionality and via a SOAP interface.