Data Security

Security for an ETL System can be simplified by establishing an overall rule that end-users cannot be allowed to access any data during the ETL process.Access to the ETL system should be limited to only those who need access to make the ETL process work. Security for the data, which is output to the data warehouse, that the ETL process supports, should be addressed as a data warehouse issue and not as part of the ETL process.

Data Profiling

This is a systematic review of the quality, scope and context of the data source from which the Extract Transform Load (ETL) system will extract data. A 'clean' data source will require minimal transformation, while a 'dirty' data source will require extensive data transformation. However, in the absence of proper planning unforeseen obstacles may arise, such as Data Profiling may reveal at the time of migration that an ETL process cannot support the 'dirty' data source. It is better to discover such issues in the planning stage rather than during implementation of the ETL system.

Archiving

ETL data should be stored at each step where a major transformation of data has occurred. In a basic data flow, this would occur after each step: extract, clean, conform, and deliver. Once the data is stored to a media, it must be determined if the stored data is to be archived and if so, for what length of time. If there are legal requirements, then it should be determined in the planning stage of the system.