In a bid to bring together two of the hottest new database management technologies – the Apache Cassandra NoSQL columnar store and graph databases – DataStax, Inc. today announced that it has acquired Aurelius LLC, developer of the popular open source TitanDB graph engine.
Datastax said the combination of the two companies will create the only unified distributed database management platform to include graph, analytics, search and in-memory capabilities. It also believes it can lick the scalability problems that have hampered broader adoption of graph databases.
DataStax, which claims 400 customers in more than 50 countries, has built its business delivering a Cassandra implementation that scales well and adapts easily to a variety of transactional use cases. Two increasingly popular applications of Cassandra are recommendation engines and fraud detection, which also happen to be the sweet spot of the graph database market, according to Martin Van Ryswyk, executive vice president of engineering at DataStax
Graph databases work by establishing relationships between data elements in a way that shortcuts the need to traverse relational tables. For example, a graph database can link together products in a catalog to various customer types to identify the likelihood that a person who buys one product will also buy another. In the same way it can look for activity patterns that are typical of fraudulent credit card use.
“Some people call graph databases whiteboard-friendly,” said Matthias Broecheler, managing partner at Aurelius. “The data model can be directly implemented in the database. You don’t have map relationships into tables.”
Titan conveniently already runs on Cassandra, “but we’re creating new products that integrate more tightly with the Datastax platform and tools and processes built around it,” Broecheler said.
The speed of graph databases has made them a popular tool for companies looking to implement recommendation engines that up-sell and cross-sell customers during a website session or while in a store. In its recent “IT Market Clock for Database Management Systems, 2014” report, Gartner wrote that “Graph analysis is possibly the single most effective competitive differentiator for organizations pursuing data-driven operations and decisions.”
TitanDB ranks second on the DB-Engines list of the most popular graph databases behind Neo4j, which has raised $45 million, including a $20 million round last month. Over the last year, demand for graph databases “has gone from a background noise to a roar,” Van Ryswyk said.
DataStax and Aurelius have worked together informally for some time, but the addition of the eight-person TitanDB engineering team will enable the development of “new products that integrate more tightly with the Datastax platform and tools and processes built around it,” Broecheler said. Aurelius claims Cisco Systems, Inc., Pearson Education, Inc. and Los Alamos National Laboratory among its customers. DataStax didn’t announce details of any forthcoming products. The company won’t sell a commercial version of Titan, which will continue to be an open source project.
The deal marks DataStax’s first acquisition, a milestone that represents the “maturation of the NoSQL market,” Van Ryswyk said. “I think we’ll see more consolidation as we’re hearing from customers that they want graph, time-series, columna, JSON and other models on one platform. That’s where we’re evolving.”