Data is stored as within data containers in the distributed file system (DFS) of the platform's data layer ("data store"). A single container can be used to store different types of data — NoSQL ("key-value"), time-series, stream data, or simple data objects (files). The data is stored as simple objects with attribute metadata, and can be grouped into collections according to the data type. For more information, see Data Objects. As explained in the data-layer overview, you can store and access container data using a variety of APIs, which accommodate multiple data formats, without duplicating the data.
All platform instances have some predefined containers, and you can create additional custom containers. The best practice is to have a dedicated container per application. You can also organize data within containers in directories (both in the predefined and custom containers). To learn how to create and delete data containers and container directories, browse their contents, and ingest and consume files, see Working with Data Containers.