Cloud Computing Today recently spoke to John Fanelli, DataTorrent’s VP of Marketing, about Big Data, real-time analytics on Hadoop, DataTorrent RTS 2.0 and the challenges specific to performing analytics on streaming Big Data sets. Fanelli commented on the market reception of DataTorrent’s flagship product DataTorrent RTS 2.0 and the mainstream adoption of Big Data technologies.
1. Cloud Computing Today: Tell us about the market landscape for real-time analytics on streaming Big Data and describe DataTorrent’s positioning within that landscape. How do you see the market for real-time analytics evolving?
John Fanelli (DataTorrent): Data is being generated today in not only unprecedented volume and variety, but also velocity. Human created data is being surpassed by automatically generated data (sensor data, mobile devices and transaction data for example) at a very rapid pace. The term we use for this is fast big data. Fast big data can provide companies with valuable business insight, but only if they act on them immediately. If they don’t, the business value declines as the data ages.
As a result of this business opportunity, streaming analytics is rapidly becoming the norm as enterprises rush to deliver differentiated offerings to generate revenue or create operational automated efficiencies to save cost. But it’s not just fast big data alone; it’s big data in general. Organizations have plenty of big data already in their Enterprise Data Warehouse (EDW) that is used to enrich and provide greater context to fast big data. Some examples of data that drives business decisions include customer information, location and purchase history.
DataTorrent is leading the way in meeting customer requirements in this market by providing extremely scalable ingestion of data from many sources at different rates (“data in motion” and “data at rest”), combined with fault tolerant, high performing analytics; flexible Java-based action and alerting, delivered in an easy to use and operate product offering, DataTorrent RTS.
The market will continue to evolve toward making analytics easier to use across the enterprise (think non-IT users), cloud-based deployments and even pre-built blueprints for “enterprise configurable” applications.
2. Cloud Computing Today: How would you describe the reception of DataTorrent RTS 2.0? What do customers like most about the product?
John Fanelli (DataTorrent):Customer feedback DataTorrent RTS 2.0 has been phenomenal. There are many aspects of the product that are getting rave reviews. I have to call out that developers have reacted very positively to the Hadoop Distributed Hast Table (HDHT) feature as it provides them with a distributed, fault-tolerant “application scratchpad,” that doesn’t require any external technology or databases. Of course, the marquee features that have the data scientist community abuzz are Project DaVinci (visual streaming application builder) and Project Michelangelo (visual data dashboard). Both enable quick experimentation over real-time data and will emerge from Private Beta over the coming months.
3. Cloud Computing Today: How would you describe the differentiation of DataTorrent RTS from Apache Spark and Apache Storm?
John Fanelli (DataTorrent):DataTorrent provides a complete enterprise-grade solution, not just an event-streaming platform. DataTorrent RTS includes an enterprise-grade platform, a broad set of pre-built operators and visual development and visualization tools. Enterprises are looking for what DataTorrent calls a SHARPS platform. SHARPS is an acronym for Scalability, Highly Availability, Performance and Security. In each of the SHARPS categories, DataTorrent RTS is superior.
4. Cloud Computing Today: What challenges do you foresee for Big Data achieving mainstream adoption in 2015?
John Fanelli (DataTorrent): Fast big data is gaining momentum! Every day I speak with customers and prospects about their fast big data, the use-case requirements and the projected business impact. The biggest challenge they share with me is that they are looking to move faster than they are able due to existing projects and technical skills on their team. DataTorrent RTS’ ease of use and operator libraries supports almost any input/output source/sink and provides pre-built analytics modules to address those challenges.