I am looking for scalable alternative to traditional DBMS like PostgreSQL or MySQL. In traditional databases I don't have the following features:

Auto sharding to ensure linear scalability. Replication with automatic failover and recovery to ensure high availability. No single point of failure. MongoDB looks like good candidate if I can sacrifice transactions.

Also I've looked at several newSQL databases. NewSQL seems suitable for my purposes: VoltDB, TiDB, cockroachDB. But I'm worried about whethever they are production-ready.

May be there are extensions allowing to run postgreSQL or MySQL in clustered mode out of box.

  • "Replication with automatic failover and recovery to ensure high availability" - Postgres can absolutely do that there are several tools that will help you to implement that. You can do sharding with built-in features in Postgres (using foreign data wrappers and inheritance) - but it's far from being "automatic". For "automatic" sharding you should probably look into Postgres-XL or maybe BDR
    – user14090
    Feb 27, 2017 at 13:12
  • Thanks, Postgres with suitable utils looks like really good approach. Feb 27, 2017 at 19:20
  • FYI: transaction support is coming in MongoDB 4.0: mongodb.com/transactions
    – Stennie
    Mar 31, 2018 at 0:57

1 Answer 1


YugaByte DB is an open-source cloud-native database for mission-critical applications. YugaByte DB supports two popular NoSQL APIs, Apache Cassandra Query Language (CQL) and Redis, in a wire-compatible fashion. It supports single-row ACID (with multi-row transactions coming soon), and has been extensively tested in a number of real-world use cases. Automatic sharding and Raft-based replication are built-in features of YugaByte DB. Please see YugaByte DB architecture documentation for more details, and see the quick start guide.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.