Amid a gust of new NoSQL database types -- key-value, graph-based, JSON-oriented or what have you -- it is possible...
to lose sight of established NoSQL software that stood as an alternative to SQL long before industry upstarts put a "No" in front of "SQL."
A noticeable example is MarkLogic. Since its inception in the early 2000s as an XML data repository, it has come to include SQL-like features and other traits, and now claims more than 425 customers, ranging from publishing houses like Condé Nast to airlines like United.
Like today's NoSQL, XML in its early days was sometimes touted as an SQL slayer. Despite considerable backing from no less than Tim Berners-Lee, inventor of the World Wide Web, XML and its related tools did not become too mainstream.
What did happen, however, is that XML found major use in the world of documents. Its ability to add meaning to documents through sophisticated encoding has helped it gain a place in government applications, content management, data integrations and other areas. MarkLogic has been at the front of the XML movement, to the point that it has outpaced all other NoSQL offerings in revenue, according to analyst group Wikibon's estimates.
"We think of MarkLogic as a document store. But 'document store' covers a lot of ground," said Joe Pasqua, senior vice president for product strategy at MarkLogic. "A document can be unstructured data, structured data or semistructured data," he said, pointing to the diversity of document approaches.
Customer requirements have led MarkLogic far beyond the basic XML roots that defined it at its inception, according to Pasqua. The MarkLogic software now has come to include integrated search, high security, geospatial querying and ACID transactions.
XML support was key to the selection of MarkLogic for work on the Founders Online website, a collaborative effort of the National Archives and the University of Virginia Press that offers the public online access to the papers of early American leaders such as George Washington, John Adams and Thomas Jefferson.
The work grew out of earlier efforts, with this project placing special emphasis on interactive performance, according to David Sewell, editorial and technical manager at ROTUNDA, The University of Virginia Press.
Performance was newly crucial because the online website would open up these important historical papers to a wide public, not just historians and other academics. MarkLogic's integrated Web server helped with fast page rendering, according to Sewell. Meanwhile, MarkLogic XQuery development tools helped improve query performance over earlier XPath query methods.
"This will support much higher user load," he said, noting that each XML file -- at this point there are almost 150,000 documents -- includes both text and descriptive metadata.
"This could be in a relational database," said Sewall, "but for us it is easier to maintain the document format. That makes it easier to index, and to do so quickly." In the future, he said, the team would look to use MarkLogic's geospatial semantic search capabilities as well. Such capabilities are somewhat akin to those of graph databases, where relationships between documents can be creatively mapped, he said.
Tale of inclusion
In its tale of inclusion, the MarkLogic Enterprise NoSQL Database could point to evolution that newer NoSQL databases may undergo. Increasingly, they will add features that enhance the original "purpose-built" architecture.
What is next for MarkLogic's database? It is JSON support, said Joe Pasqua.
That is ironic in a way, because JSON began life as something of an alternative to XML. While XML found favor in document-centric processing, its reception in general programming was less favorable.
XML was seen as a verbose and overhead-laden format by many software developers working in Web applications. That created an opening for JSON.
So, MarkLogic's upcoming JSON support is really just part of a larger Web application trend. Continued growth of the JSON format has even led major SQL houses like IBM, Microsoft, Oracle and Teradata to support JSON data formats in their flagship databases.
JSON is at the heart of up-and-coming document-oriented NoSQL databases from players like MongoDB and Couchbase, as well. JSON support is due this fall with MarkLogic Version 8, Pasqua said.
Investigate NoSQL roots
Discover NoSQL use cases
View an expert video comparing SQL and NoSQL