I have started to learn about "No SQL" not because Google (bigtable) or Facebook using "NO SQL", just because Oracle has announced its No SQL DB (here and here) which uses Key-Value Data model based on its "Oracle Berkeley DB Java" (here and here).
Based on "Martin Fowler" (YouTube, Brief PDF) "No SQL" have not a unique definition and we should define it with its characteristics which are:
- Mostly Open Source
- Developed in 21st Century and Web
- Key-Value ( i.e. Project Voldemort, Riak, Redis)
- Document (i.e. mongoDB, Raven DB, CouchDB)
- Column-Family (i.e. Apache HBASE, Cassandra)
- Graph (i.e. Neo4j)