Whether we realize it or not, we are living in an era of big data. Just to get an idea about how much data is constantly being generated, just look into some of the numbers published by Google for 2019. As we know, Google Photos is the multimedia repository for storing photos created by Google. In 2019, an average of 1.2 billion photos and videos were uploaded to Google Photos every day. Also, an average of 400 hours of video (amounting to 1 PB of data) were uploaded every minute each day to YouTube. We can safely say the amount of data that is being generated has simply exploded.

The current interest in data-driven algorithms is driven by the fact that data contains valuable information and patterns. If used in the right way, data can become the basis of policy-making decisions, marketing, governance, and trend analysis.

For obvious reasons, algorithms that deal with data are becoming more and more important. Designing algorithms that can process data is an active area of research. There is no doubt that exploring the best ways to use data to provide some quantifiable benefit is the focus of various organizations, businesses, and governments all over the world. But data in its raw form is seldom useful. To mine the information from the raw data, it needs to be processed, prepared, and analyzed.

For that, we first need to store it somewhere. Efficient methodologies to store the data are becoming more and more important. Note that due to the physical storage limitations of single-node systems, big data can only be stored in distributed storage consisting of more than one node connected by high-speed communication links. So, it makes sense that, for learning data algorithms, we start by looking at different data storage algorithms. 

First, let's classify data into various categories.