DataStores

A datastore is a storage system used to persist digital data for applications and services. It provides mechanisms to store, retrieve, and manage data across different formats and access patterns. The term is often used broadly to describe any repository of data, but it is typically contrasted with a database: a datastore may emphasize scalable storage and retrieval, while a database also provides structured querying and transactional guarantees. In practice, a datastore can encompass databases, object stores, and other storage engines.

Common datastore types include relational databases (SQL) for structured data and strong consistency; document stores that model data as semi-structured documents (JSON, BSON); key-value stores optimized for simple lookups; wide-column stores that organize data by rows and column families; and graph databases for connected data. Additional datastores include object stores and file stores used to hold unstructured data. Some systems combine features to support multiple data models.

Key considerations when selecting a datastore include data model, query capabilities, consistency and durability guarantees, scalability, latency, and throughput. Other factors are security, access control, backup and recovery, and governance. Cloud providers often offer managed datastore services with automated replication, versioning, and monitoring, enabling developers to focus on application logic rather than storage infrastructure.

Datastores underpin a wide range of applications, from user session data and product catalogs to analytics, graphs of social networks, and logs. Understanding the data access patterns and durability requirements helps guide the choice of datastore and its configuration.