MemSQL 3.1

MemSQL's latest release takes the world's fastest database platform to the next level with major new features and performance improvements. MemSQL 3.1 introduces:

  • Automatic cross-datacenter replication, provisioned with just a single command.
  • Support for views as well as table, column, and row level security.
  • Highly performant, transactional bulk data loading from a filesystem or from Amazon S3.
  • Performance improvements in the flash-optimized compressed column store.

MemSQL 3.1 is available immediately for download

Click the button below to download MemSQL 3.1. Visit our documentation at for an in-depth list of new features and resources on how to get up & running with MemSQL quickly.

Download Now

MemSQL v3.1 in Action.

Learn how MemSQL helped CPXi eliminate ETL and streamline their real-time bidding infrastructure.

  Read the Case Study

MemSQL 3.0

MemSQL now features an integrated tiered storage architecture which consists of a transactional in-memory row store and a highly compressed flash or disk-based column store. With MemSQL v3.0, users are no longer limited to just memory, allowing expansion of the MemSQL platform to hundreds of terabytes.

  • Tiered architecture accessible through a single interface.
  • Workloads can access data in both row and column stores.
  • Highly compressed, deployed on memory, flash, or disk.

MemSQL 2.5

JSON Data Type

MemSQL now supports the JSON data type, making it simple for developers to store and query their JSON data alongside structured data – all within one consolidated RDBMS. Learn more.

  • Analyze structured and semi-structured data together to produce a complete view into your business operations.
  • MemSQL supports standard SQL functions with JSON, including built-ins, GROUP BY, JOINs, and more.

Improved Index and Scan Performance

MemSQL performs 3x faster over large datasets, enabling scans of over 30 million rows per second per thread.


MemSQL now supports online ALTER TABLE operations (ADD/REMOVE COLUMN and ADD INDEX commands) in the distributed architecture without disturbing read and write workloads.

  • The table is available for reads and writes as the online ALTER TABLE command runs.
  • The memory overhead of running an ALTER TABLE operation is only a few megabytes regardless of the size of the table.

Online Backup / Restore

Distributed backup enables you to take a point-in-time snapshot of your data across the cluster with just a single command, while your read and write workload continues to run.

  • The RESTORE DATABASE command enables you to restore an entire distributed database onto the cluster with just a single command.
  • Backups can be restored to a cluster with a different number of nodes and/or a different redundancy level from the original cluster.

Additional SQL Support

  • Scalar subselects, subselects with EXISTS and IN.
  • Constants are parameterized in projections, as well as ORDER BY and GROUP BY clauses.
  • UNION ALL, GROUP_CONCAT, various built-in functions, and expression performance improvements.
  • Persisted computed columns on arbitrary SQL expressions.

All of these are supported on both standalone and distributed tables.