ETL - extract, transform, load
ETL architecture is used in order to manage and optimize the data together with the process of its transfer which, as a result, leads to simplification and standardization of data storage in the data warehouse.
Purpose of the ETL tool is to create a universal framework for different processes of transforming data into target destination.
The ETL letters are the abbreviations for extraction, transformation and loading of data, which means that the ETL tools extract data from a source, transform it into new formats and load it into target data warehouse.
System’s architects created a scheme which eliminates repeatedly appearing actions. Loading the data into warehouse requires two types of actions: one, in which data is transformed from the identifiable source system according to source-specific rules – this action results in transforming data into a standardized format. Each action has to be done by individual method for every data storage system.
The other, conformed type of ETL architecture means creating universal way of data entity, regardless of the source and purpose; it allows to follow reusable rules which are applied in business and work in different conditions with variety of data. Both systems has pros and cons. At the beginning the first one requires fewer actions to be performed in order to work smoothly and seems easier. However, looking at the overall life cycle of the data warehouse, in the long run using the conformed ETL is more efficient.
It is because data warehouse is being developed and has been growing continually when the new data is added to the system, so when the traditional architecture is taken into consideration, though it may seem easier to make it, in the long run it turns out that the same or similar actions are repeated, needlessly lengthening the development time and increasing its complexity. Implementing the universal for one entity ELT architecture requires more work than doing the same for a particular task, however in the long run it brings more benefits as the data is stored and the scheme may be used further.
Thus the most obvious advantage of using ETL is that the storage data is easily reused which results in improving its quality. Moreover it’s easier to make and use, because in fact it is less complex than storing data in different traditional systems. Additional advantage is also that it’s less difficult to add and acquire new source systems. One should also bear in mind that the conform ETL architecture is not to be applied everywhere, it should be proceeded by detailed analysis, which would ensure the benefits that would possibly result from following the conform architecture. Otherwise it may not be so beneficial.
This data architecture enables direct access to data in operational systems, which is very important and in fact shorten the time considerably, which otherwise would be spent on data implementation and simplify the data usage.