BabuDB is an embedded non-relational database system suitable for storing large numbers of arbitrary key-value pairs. It is capable of accomodating databases that exceed the main memory of the system, supports fast and computationally inexpensive snapshots, and offers prefix and range lookup functionality. To provide for failure tolerance, it also supports crash recovery and replication mechanisms.

Research perspectives include automatic partitioning of indices across multiple nodes as well as self-management mechanisms.

BabuDB has originally been developed as a storage backend for the XtreemFS metadata server. It has also proven useful for other application scenarios, such as web content storage.