Xception

DBenchDW

Data Warehousing for the Analysis and Sharing of Results from Dependability Evaluation Experiments

The approach is to collect the raw data produced in dependability evaluation or benchmarking experiments and store it in a multidimensional data structure (data warehouse). The data analysis is done through the use of commercially available OLAP (On-Line Analytical Processing) tools such as the ones traditionally used in business decision support analysis. Current implementation uses an OracleÒ 9i database to store the data and the Web DiscovererÒ tool analyse the results through the Web.

The advantages of this approach are:

General approach
- Applicable to all the experimental dependability evaluation scenarios
- Existing tools and experimental setups can be used without changes
- Use standard OLAP technologies and tools for the analysis of multidimensional datasets
Makes it easy to compare and cross-exploit raw results from different experiments, as all the raw data is stored in a common data warehouse.
Makes it easy to share raw results world wide as the data stored in a data warehouse can be explored by web-enabled versions of OLAP tools.

Links