YugabyteDB

YugabyteDB is a distributed SQL database compatible with PostgreSQL, built to stay up across zones and regions; open source, or managed as YugabyteDB Aeon.

YugabyteDB keeps a relational database online when a server, a zone or an entire cloud region fails, by spreading every table across a cluster of nodes. Yugabyte, the Sunnyvale, California company founded in 2016, develops it. The database speaks the PostgreSQL dialect and loads PostgreSQL extensions, so code written for PostgreSQL runs against it. Its core is public under the Apache 2.0 license, and the same engine sits behind the managed and self-managed editions. Two query layers share one storage engine. YSQL is the PostgreSQL-compatible API and YCQL the Cassandra-compatible one, and both write to DocDB. DocDB splits tables into tablets and replicates each one synchronously through Raft, so a follower takes over when a leader fails. Read replicas and xCluster replication copy data asynchronously to other regions, while row-level geo-partitioning pins chosen rows to the region where they must stay. Change data capture follows the PostgreSQL logical replication model. Query Plan Management keeps plan history so a known good plan can be restored, and Performance Advisor watches workloads in YugabyteDB Aeon. The database is aimed at engineering and platform teams whose services must keep accepting writes through an outage, and at teams moving PostgreSQL workloads onto distributed SQL. On YugabyteDB Aeon, the managed service, startups and scale-ups that need zone-level resilience start on the Standard tier. Global businesses that need multi-region topologies or data residency move to Professional. Enterprises that need single sign-on, customer-managed keys and disaster recovery workflows take Enterprise, which can also run inside their own cloud account. Teams that prefer to operate it themselves use the open source build or YugabyteDB Anywhere. Yugabyte points the database at database and application modernization, cloud-native services, edge and IoT systems, and generative AI apps that keep embeddings in the pgvector extension. Platform engineers provision Aeon clusters as code through the Terraform provider, the REST API or the CLI. Row changes leave through change data capture to external consumers, such as Debezium connectors feeding Kafka topics. Enterprise deployments add integrations with Commvault, Delphix and Nutanix for the data protection and infrastructure they already run. The fit is strongest for a customer-facing service spread over several regions, where losing one region must not stop the rest.