I have worked with several companies that initially started working with their OLTP to provide analytics. This will generally work early on, but these systems aren’t optimized to run complex queries.
Databases like MongoDB and CassandraDB also add in the complexity of not being SQL-friendly, which most analysts and data practitioners are accustomed to using. This further makes these data systems a poor choice in the long run for performing analytics.
Why does it really matter what type of data storage system you use?
Data is data, and if you want to ask a question to your transactions database, you should be able to, right?
Yet we generally distinguish between OLTP and OLAP.
Also known as online transaction processing vs. online analytical processing, both are different use cases that benefit from contrasting designs and underlying software. This is why solutions like Teradata and Vertica were massive back in the day(and honestly still play a large role in many enterprises). Teradata was actually one of the first data warehouse systems to manage a TB of data for Walmart.
But the question becomes, why?
Why do we need to duplicate our date and remodel it for analytics?
In this article we will discuss some of the reasons why you will need to eventually develop an analytical system to answer your business questions.
Access Patterns
Keep reading with a 7-day free trial
Subscribe to SeattleDataGuy’s Newsletter to keep reading this post and get 7 days of free access to the full post archives.