A new version of the RavenDB document database has been released. The improvements are mainly to performance, and the team has fixed the problems with memory use in the earlier release.

RavenDB is a schema-free database from Hibernating Rhinos that is open-source with a commerical option.

It supports sharding, replication and multi-tenancy and comes with a client API for .NET, Silverlight, Javascript and REST over HTTP.  RavenDB is built using .NET, and integrates well with the .NET environment.

Data in RavenDB is stored schema-less as JSON documents, and can be queried efficiently using Linq queries from .NET code or using RESTful APIs. Internally, RavenDB make use of indexes which are automatically created based on usage, or are created explicitly by the consumer.

In the new version indexes have also been added for referenced documents, and indexes are defined using Map/Reduce functions written in Linq syntax. New profiling options have been added, so you can now choose a cloud or
production profile in addition to the existing options such as web and embedded. Raven is built on Lucerne, and you can make use of Lucerne’s built-in features.

This release has taken six months of development, but the team of developers at Hibernating Rhinos now plan on bringing out an update every six to eight weeks. The list of planned improvements include robust SQL replication, automatic index merges, and client support for WinRT and Mono for Android.