Despite the advances we’ve made in data science and advanced analytics in recent years, many projects still are beholden to a technological holdover from the 1980s: extract, transform, and load, or ETL. It’s uncanny how those three letters strike fear into the hearts of data architects, but we seem powerless to move beyond it. Is there anything that can save us from the madness of ETL?
Before looking at potential successors to ETL, let’s look at the origins of the technology. As companies amassed ever-bigger amounts of transactional data in their production databases in the 1980s and 1990s, they realized they needed dedicated business intelligence (BI) systems for analysis and reporting. In many ways, BI put the “p” back into enterprise resource planning (ERP).
Author: Alex Woodie