Once upon, a database program was someone who knew SQL. Today, they need to know one of SQL, JavaScript, XQuery, JSONIQ, SPARQL, Tomorrow, they'll need to know all of these and perhaps none of them. What is most significant about the NoSQL movement is that it has raised the awareness of non-RDBMS applications, but in the process has also raised an intriguing question. Will there come a time when a database is simply a place to store and retrieve data, and ultimately the characteristics of that data are no longer the responsibility of the application developer or database administrator, but the analyst or end-user? Increasingly, the answer is yes, and sooner than most would admit.

Database vendors, seeking to be competitive, are going back to the index and rethinking what being a database really means. Hadoop Yarn Containers, MarkLogic Server, JackRabbit and other data systems are emerging that integrate JSON, XML, relational and semantic triple capability into their core architecture, along with a "data application" layer that is increasingly intelligent at being able to store and optimize this data for the widest potential needs. Tools like JSONIQ provide data abstraction layers to do similar things at the query layer. Distributed datastores and memcached in-memory systems provide the scalability and performance respectively to make such systems work at or near real time.

This session looks at the evolution of the virtual database and it how it will influence data management in the next decade.