Valuation Factors for the Necessity of Data Persistence in Enterprise Data Warehouses on In-Memory Databases

ETL (extraction, transformation, and loading) and data staging processes in Enterprise Data Warehouses have always been critical due to their consumption of time and resources. Mostly, the staging processes are accompanied with persistent storage of transformed data to enable a reasonable performance when accessing for analysis and other purposes. The persistence of – often redundant – data requires high effort regarding maintenance, such as for updating and for guaranteeing consistency. Concurrently, flexibility of data usage and speed of data availability is delimited. Especially column-based inmemory databases (IMDB) enable to query high data volumes with good response times. Progress in in-memory technology leads to the question how much persistence is necessary in such warehouses. The answer cannot only be based on cost models, but also has to take other aspects into account. The presented thesis project deals with the problem of defining valuation factors for supporting the decision whether to store data in Enterprise Data Warehouses based on in-memory databases.