Microsoft researchers: NoSQL needs standardization

The ever-growing number of non-relational, or NoSQL, databases needs standardization in order to thrive, two Microsoft researchers argue in the new issue of the Association for Computing Machinery's flagship publication, Communications.

"The nascent NoSQL market is extremely fragmented, with many competing vendors and technologies. Programming, deploying, and managing NoSQL solutions requires specialized and low-level knowledge that does not easily carry over from one vendor's product to another," the two researchers, Erik Meijer and Gavin Bierman, published in the April issue of Communications.

The pair of researchers offer a mathematical data model and standardized query language that could be used to unify NoSQL and SQL data models, work they call "coSQL."

"There is little to disagree with in this paper," said James Phillips, a co-founder and vice president of products for NoSQL database vendor Couchbase, who had no involvement in the work. "I firmly support the conclusion that a standardized data manipulation language would accelerate market adoption of NoSQL database technologies by eliminating developer-impacting fragmentation."

Over the past few years, , including CouchDB, Cassandra and MongoDB. Administrators have found these new data stores more suitable than relational databases for tasks such as storing large amounts of data across multiple servers, or for easily storing information that does not need to be indexed for complex querying.

Meijer and Bierman compare this current flourish of non-relational databases to the proliferation of relational databases in the early 1970s. At that time, developers would have to understand the peculiarities of each database, as well as how to interact with the underlying hardware. What unified this industry was the widespread adoption of SQL (Structured English Query Language), the researchers argue.