DataTorrent recently announced the general availability of DataTorrent Real-Time Streaming, a platform that delivers real-time analytic capabilities on Apache Hadoop that allow users to obtain actionable business intelligence from streams of Hadoop data. DataTorrent Real-Time Streaming boasts the ability to run analytics on streams of Hadoop data at volumes of over 1 billion events per second by using in-memory processing with low to zero latency. Whereas comparable technologies such as Spark Streaming from Apache Spark split a stream of Hadoop data into segments and performs in-memory processing, DataTorrent Real-Time Streaming operates directly on Hadoop containers without scheduling batches of Hadoop streams for processing. By avoiding the scheduling overhead associated with processing “mini-batches” of Hadoop data, DataTorrent claims operational efficiencies that allow it to process more Hadoop events with sub-second latency than competing products.
Phu Hoang, co-founder and CEO, DataTorrent, remarked on the innovation enabled by DataTorrent Real-Time Streaming as follows:
Hadoop has made big data analytics a reality; however, the true value of big data is unlocked when it can be acted upon in real-time. DataTorrent Real-Time Streaming is designed specifically to address this need for the enterprise. Through the advances provided by Hadoop 2.0, we are proud to raise the bar on real-time analytics to offer the industry’s first true real-time data ingestion and analysis platform at scale.
Designed specifically for Hadoop 2.0 and the enhancements enabled by YARN, DataTorrent RTS also boasts the ability to perform complex, high performance computation on streaming Hadoop data with high availability. Certified to work with Hadoop distributions from Cloudera, Hortonworks and MapR, DataTorrent RTS represents a commercial product that plays in the increasingly hot space constituted by products intended for real-time analytics on streaming Big Data alongside the likes of Apache Storm, Apache Spark and Amazon Kinesis. Questions of performance aside, one of the keys to DataTorrent’s success will be its ease of implementation and ability to simplify and streamline the derivation of meaningful analytics from streaming Hadoop data. To date, the Santa Clara-based company has raised $8M in Series A funding in a round led by August Capital.