What are some NewSql relational scalable databases


Relational database management with the desired scalable performance from NoSQL by combining OLTP and ACID schemes

NewSQL is a class of relational database management systems designed to provide the scalability of NoSQL systems for OLTP (Online Transaction Processing) workloads while maintaining the ACID guarantees of a conventional database system.

Many enterprise systems that handle high-quality data (such as finance and order processing systems) are too large for traditional relational databases, but have transaction and consistency requirements that are impractical for NoSQL systems. The only options available to these organizations so far have been either to buy more powerful computers or to develop custom middleware that distributes requests across traditional DBMS. Both approaches have high infrastructure costs and / or development costs. NewSQL systems try to reconcile the conflicts.


The term was first used by Matthew Aslett, 451 Group analyst, in a 2011 research report that discussed the rise of a new generation of database management systems. One of the first NewSQL systems was the parallel H-Store database system.


Typical applications are characterized by high OLTP transaction volumes. OLTP transactions;

  • are short-lived (i.e. no user booths)
  • Touch small amounts of data per transaction
  • Use indexed lookups (not table scans).
  • have a small number of forms (a small number of queries with different arguments).

However, some support HTAP (Hybrid Transactional / Analytical Processing) applications. Such systems improve performance and scalability by eliminating heavyweight recovery or control of parallelism.


The two common differentiators of NewSQL database solutions are that they support the online scalability of NoSQL databases and the relational data model (including ACID consistency) using SQL as the primary interface.

NewSQL systems can be loosely divided into three categories:

New architectures

NewSQL systems use different internal architectures. Some systems use a cluster of shared nothing nodes in which each node manages a subset of the data. This includes components such as distributed parallelism control, flow control and distributed query processing.

SQL engines

The second category is optimized storage engines for SQL. These systems provide the same programming interface as SQL, but are more scalable than integrated engines.

Transparent shard

These systems automatically split databases across multiple nodes using the Raft or Paxos consensus algorithm.

See also