The evolution of database management systems is often simplified into a linear progression from SQL to NoSQL and then to NewSQL. However, historical development was non-linear, as distributed relational databases existed long before the term NewSQL was coined.
Traditional relational databases focused on vertical scaling for decades. As internet services grew, the need for horizontal scaling emerged, leading to the rise of NoSQL. These systems varied significantly in data models, interfaces, and consistency guarantees rather than following a single standard.
Analyst Matthew Aslett introduced the term NewSQL in 2011 to describe high-performance SQL systems. Later, researchers like Andy Pavlo and Matthew Aslett narrowed the definition to modern relational databases that combine ACID compliance and SQL with horizontal scalability.
The term Distributed SQL, which gained prominence alongside NewSQL, focuses on architecture where data is physically partitioned across multiple nodes. Unlike NewSQL, this term emphasizes the system's inherent design for distributed storage, replication, and coordination.
Modern database evaluation requires looking beyond category labels. Architects should analyze data distribution methods, transaction support, replication mechanisms, and topology, such as whether a system is designed for single-region or multi-region deployment.
