"We have seen an explosion in front-end tools for data discovery, and YarcData's mission is to be the server back-end for data discovery, similar to how MPP appliances were the server back-end for traditional business intelligence", stated Arvind Parthasarathi, President of YarcData. "With our new software release, the Urika appliance further accelerates the uncovering of valuable insight in disparate enterprise data by combining scalable performance and industry-standard interfaces with the existing analytic ecosystem. Urika also eliminates the need for extensive data preparation, modelling and knowing all of the questions to be asked upfront."
One of the central promises of Big Data is the ability to discover new insights and unknown relationships in the data. This poses a fundamental challenge for many traditional analytics tools, since the discovery process demands the ability to ask questions in an ad hoc, iterative fashion, to add new data sources on the fly as required, and to do all of this without modelling the data beforehand.
"Urika enables us to deliver on our mandate to deliver Big Data analytics to our world class researcher base looking for breakthrough discoveries", stated Nick Nystrom, Director, Strategic Applications at the Pittsburgh Supercomputing Center. "Researchers come to us seeking to discover unknown, hidden relationships in their data and they rely on Urika's real-time response to their most complex queries on their largest datasets - allowing them to explore hundreds of hypotheses in the time previously taken to explore just one."
With this new software release, Urika now integrates with a broad array of enterprise interfaces, including W3C industry standards SPARQL and RDF, JDBC (Java DataBase Connectivity), JSON (Java Script Object Notation), and Apache Jena. Additionally, analysts and data scientists can now easily interact with Urika using a wide variety of current and emerging third-party visualization and BI tools including Centrifuge Visual Network Analytics, and TIBCO Spotfire. This allows enterprises to deliver powerful new discovery analytics capabilities to business users while retaining the familiar user experience of their existing front-end tools. YarcData has also formed partnerships with a variety of data discovery ecosystem providers, such as Cloudera, Centrifuge, and TIBCO to explore deeper integration and deliver a more seamless data discovery experience.
Targeted performance improvements in the SPARQL query engine now enable Urika to handle key operations on aggregate functions up to 400 times faster, further advancing Urika's existing orders-of-magnitude performance advantage. In addition, a significant improvement in memory efficiency enables analysts to load even larger data sets and simultaneously run complex analytical queries to rapidly investigate multiple changing hypotheses.
In contrast with traditional analytics where questions are typically defined and fixed up front, data discovery is an iterative process of hypotheses validation. This means analysts need to track and build upon previous steps in their hypothesis creation. With this new release, the hypothesis validation monitor capability provides users with fine grained detail across the lifecycle of hypotheses and their validation results. Analysts can re-use queries, analyse performance and investigate query results to develop and expand an existing hypothesis or series of hypotheses in order to get to important insights faster.