Forrester Research has published a white paper Internet Of Things Applications Hunger For Hadoop And Real-Time Analytics In The Cloud ($US499) in which it argues that Hadoop and streaming analytics platforms are key components of IoT data-driven applications.
According to Forrester enterprise applications must handle IoT data in two ways: analyse large volumes to find patterns and insights that can be valuable in the future; and perform streaming analytics to glean immediate, actionable insights.
“As firms clamor to find actionable insights in IoT data, they must find an economical landing zone for these always big and sometimes gnarly data sources. Hadoop combined with streaming analytics is it,” Forrester says. “Hadoop can store and process large volumes of data across a linearly scalable cluster. Streaming analytics platforms can sense and respond in real time.”
Forrester says Hadoop is ideal for analytics at rest and that a number of new service analytics services have emerged in recent years, all cloud-based. “AWS launched Amazon Kinesis in late 2013, Google announced Google Cloud Dataflow in 2014, and Microsoft Azure Streaming Analytics is in customer preview.”
It adds: “Perhaps the most interesting options are streaming analytics platforms that can run directly on a Hadoop cluster. Named applications include Data Torrent, Apache Spark, Apache Samza and Apache Storm. It has more information available in The Forrester Wave: Big Data Streaming Analytics Platforms, Q3 2014 ($US2495).