A shared nothing, document-oriented cluster data store. Accessed via SQL and has builtin BLOB support. Uses the cluster state implementation and node discovery of Elasticsearch. License: Apache 2.0, Query Method: SQL, Clients: HTTP (REST), Python, Java (JDBC or native), Ruby, JS, Erlang, Replication + Sharding: automatic and configurable, written in: Java, » Crate Data GitHub Project, » Documentation.
API: Scala. Written in Scala. Replication: Replicas vote on writes and reads. Sharding:Hashes keys onto array of replica cohorts. Concurrency: Optimistic + Multiversion Concurrency Control. Provides multirow atomic writes. Exposes optimistic concurrency through API to support HTTP Etags. Embedded solution.