“All of the same features and functions we have for XML data are available for data that comes in from JSON,” said Joe Pasqua, MarkLogic executive vice president for products.
A precursor to the wave of NoSQL database servers that hit the market about six years ago, MarkLogic Server was introduced in 2005 as an alternative to standard SQL-based relational databases. The software used the then much-hyped XML format to store and deliver records.
To use MarkLogic, developers and database administrators had to learn the XML query language, called XQuery, a requirement that might have hampered wider adoption among shops with traditional SQL expertise. Nonetheless, the software was adopted for a number of large projects—such as helping to power the U.S. government’s healthcare.gov site and providing asset management needs for publishing giants Elsevier and Conde Nast.
“It just doesn’t make sense to slurp out all of the data from the database to a middle tier and do all of the processing there,” Pasqua said. “It makes more sense to take a smaller amount of code, move it into the server, and have it operate there where it’s close to the data.”
MarkLogic isn’t the only database system to catch JSON fever—most traditional enterprise database systems now support the format in some way. The open source PostgreSQL supports the protocol as a native data type, and MongoDB was built from the ground up to use JSON as its primary format.
By offering the full set of processing capabilities it developed for XML, MarkLogic hopes to have an edge in the market.
Also new for version 8: MarkLogic can make more inferences about a query, returning results that may be relevant but don’t include the actual query terms. That’s thanks to the new support of Sparql, a query language for rich Semantic web documents.
MarkLogic 8 starts at US$8,500 per year for in-house use or $0.99 per hour for a cloud version hosted by Amazon Web Services. As an enticement for new customers, the company is offering free use of an AWS-hosted version for one year.