|#The key differences between RDBMS and NoSQL databases:|
For the last 20 or 30 years, classic data warehousing has been based on the same regimented approach. However, the future is changing this. Traditionally, processes such as identifying data lineage, documenting metadata, and being able to reconcile data across different reports coming from different data tables in different data marts have been critical to ensure that numbers are correct.
- This standard approach to data warehousing has been important, in order to have confidence in your data and meet regulatory and compliance requirements.
- However, over time businesses have become more complex and are doing things at an ever-accelerating pace, and as a result, data storage needs for analytics are changing.
- For example, there was a time when the core business of a retail establishment was to understand and sell a single line of products through a bricks-and-mortar presence.
However, within the last 5 to10 years, that has evolved into retail establishments needing to sell and fulfill their products through multiple online, offline, and mobile channels, while understanding the dynamic way consumers research and choose to purchase a retail product. Similar issues also exist in the telecommunications industry.
- In the early days of the industry, companies sold fixed lines, and the only product flavors were local, long-distance, and international calls.
- Today, companies must deal with mobile, data over mobile, data over different bandwidths, and all of the product permutations that people have, not to mention the increased competition. Companies also must track service switches and pricing tiers.
More: MongoDB 3.2 fundamentals for Developers-Learn with Exercises
The relational databases of the past will work alongside less formally structured database schemas, such as NoSQL, that do not require structured and logical data but lend themselves well to the fast retrieval and analysis of multiple data types. These analytical warehouses will be much more flexible and dynamic, lending themselves to adhoc analyses and to bringing multiple disparate data sources together in order to address critical business questions.
This will be essential in the future, as companies do things with data such as combine Web behavior click-stream data with Twitter feeds, customer satisfaction surveys, product purchase records, and third-party credit data, in order to use data science techniques to look for relationships.