CITUS DATA, owned by Microsoft, published distributed CITUS 13.0 DBMS implemented in the form of an extension to PostgreSQL 17. CITUS provides horizontal scaling of PostgreSQL in a cluster based on typical equipment and allows data to be spread across nodes using sharding with settings at the column and storage scheme levels. For applications, the CITUS cluster appears as one large PostgreSQL server uniting the resources of its components. The code is written in the language of SI and is distributed under the AGPLv3 license.
Sharding enables the storage organization of a very large amount of data, exceeding the local drives of individual cluster components. Replication across multiple nodes provides fault tolerance and ensures operational continuity in case of component failures. Customer requests are distributed among servers to ensure predictable query execution times on large, real-time data sets.
Changes in version 13.0 of CITUS :
- Transition to using PostgreSQL 17.
- Ability to use the Json_Table() function in distributed queries to convert JSON data into a relational format compatible with PostgreSQL tables.
- Support for the “Merge … when not Matched by Source” syntax.
- Option to set access method for distributed partitioned tables using “Create Table … Using” and alter the access method with “Alter Table .. Set Access Method.”
- Addition of support for automatically generated identity columns in distributed partitioned tables.
- Introduction of support for using exclusion constraints in distributed partitioned tables.
- Resolution of issues with role synchronization between nodes.
- Enhanced data distribution for sharding