Until recently, ETL involved uploading data at regular (i.e., monthly or weekly) time intervals to drive business performance decisions and identify business opportunities. However, as BI tools become more integrated with overall business functions, including business performance management (BPM) and reporting and analysis requirements, data needs have shifted from monthly or weekly intervals to real time updates. This means that it has become more important for data transfers to accurately reflect real time business transactions, and that there has been an increase in the amount of data transfers required.
Nonetheless, real time ETL doesn't necessarily refer to automatic data transfer as operational databases are updated. In terms of BI, real time may mean different things to different organizations or even different departments within these organizations. Take, for instance, an automotive manufacturer whose traditional data warehouse solutions (OLAP cubes, etc.) involved capturing data at a given point in time. The automotive manufacturer might, for example, have wanted to track and compare monthly sales with last year's sales during the same month by region, car model, and dealer size, thus requiring the data warehouse to be updated on a monthly basis. However, as the manufacturer's business decisions evolved based on this analysis, its data needs shifted from a monthly requirement to a weekly one, and on to an ever more frequent basis, eventually creating the demand for real time data. In the case of the automotive manufacturer, real time data may be useful for identifying the movement of car parts within a warehouse relative to their storage locations and comparing this information with the demand for these parts.
Such a shift in data requirements affects both the volume of data required and when the data loading occurs. The end result is that, in order to meet the changing needs of user organizations, ETL and BI vendors have concentrated on moving towards real time ETL and shifting their data loading functionality to accommodate higher volumes of data transfer.
Nonetheless, real time ETL doesn't necessarily refer to automatic data transfer as operational databases are updated. In terms of BI, real time may mean different things to different organizations or even different departments within these organizations. Take, for instance, an automotive manufacturer whose traditional data warehouse solutions (OLAP cubes, etc.) involved capturing data at a given point in time. The automotive manufacturer might, for example, have wanted to track and compare monthly sales with last year's sales during the same month by region, car model, and dealer size, thus requiring the data warehouse to be updated on a monthly basis. However, as the manufacturer's business decisions evolved based on this analysis, its data needs shifted from a monthly requirement to a weekly one, and on to an ever more frequent basis, eventually creating the demand for real time data. In the case of the automotive manufacturer, real time data may be useful for identifying the movement of car parts within a warehouse relative to their storage locations and comparing this information with the demand for these parts.
Such a shift in data requirements affects both the volume of data required and when the data loading occurs. The end result is that, in order to meet the changing needs of user organizations, ETL and BI vendors have concentrated on moving towards real time ETL and shifting their data loading functionality to accommodate higher volumes of data transfer.
No comments:
Post a Comment